Chemically Amplified Resist Research Articles

HSQ hybrid lithography for 20 nm CMOS devices development

This paper reports the lithographic development on 8 in. silicon wafers for 20 nm CMOS devices using a hydrogen silsesquioxane (HSQ) hybrid lithography process. The method relies on splitting the pattern into high-resolution and low-resolution components at the same level. Geometries lower than 40 nm are exposed with HSQ ebeam process, geometries between 40 and 250 nm with a chemical amplified resist (CAR) ebeam process and the remaining with a CAR deep UV process. Using this technique, we can produced 20 nm structures with an exposure time drastically lower than using only a single HSQ ebeam lithography process. Finally, a pattern transfer capability of a 20 nm isolated line into a polysilicon layer is presented.

Resist distribution effect of spin coating

The thin-film formation of the spin coating is one of the important factors in the fabrication of microelectronic devices. In this study, the theoretical models for thickness variation during spin coating and nanotopography impact are analyzed. The finite-difference–time-domain method and the finite-element method are used to solve the convective diffusion equation for solvent distribution and the Navier–Stokes equation including solvent evaporation for the film thickness change. These numerical calculations are in good agreement with experimental results for 193 nm chemically amplified resist (CAR) and i-line non-CAR resists. Solvent distributions of nonspin coating are described through mesoscale modeling by using the Monte Carlo method. Nanotopography impact on the variation of resist distribution after spin coating is investigated quantitatively. The reason for the similarity in the transfer functions for different types of wafers is due to solvent diffusion and evaporation.

Line Edge Roughness and Development Rate Analysis Based on Random Reaction/Percolation Model for Chemically Amplified Resists.

Line edge roughness (LER) in chemically amplified resists (CARs) is analyzed as a fluctuation in acid catalyzed reaction to determine molecular solubility and developer percolation. Two probability processes to cause LER are modeled: the local acid generation l diffusion process and the main reaction/ developer percolation processes. LER caused by fluctuation in the main reaction is found to be significantly larger than molecular size. Dependence of LER on various parameters such as acid concentration, diffusion length, molecular sizes, protection ratio and its variation is discussed. Our results suggested a trade-off relation between the acid fluctuation component and the main reaction component of LER in most ArF resists today. High contrast exposure characteristics of the dissolution rate in CARs today are also explained with the model.

Study of the synthesis and characterization of methacrylate photoresists

AbstractPolymethacrylate copolymers used as single‐layer resists (SLR) were investigated in this study. A photoacid generator (PAG) in the chemically amplified resist (CAR) can generate a Brönsted proton after light stimulation, which can catalyze the deprotection reaction of t‐butyl methacrylate during the postexposure baking (PEB) process. Excimer lasers (KrF and ArF) were used as light source to evaluate the lithographic performance of resists. The hydrophobic property of the resist changed to hydrophilic property after the deprotection reaction by light exposure, which made the exposed resist soluble in the base developer. The deprotection reaction was analyzed by using FTIR and the reaction kinetics of the resist is discussed. The effects of the t‐butyl methacrylate and the isobornyl methacrylate in the resists on the light sensitivity, contrast, reactive ion etching (RIE) resistance, and lithographic performance were systematically investigated. © 2002 Wiley Periodicals, Inc. J Appl Polym Sci 83: 1860–1869, 2002

Chemical and physical aspects of the post-exposure baking process used for positive-tone chemically amplified resists

Chemically amplified (CA) resists are in widespread use for the fabrication of leading-edge microelectronic devices, and it is anticipated that they will see use well into the future. The refinement and optimization of these materials to allow routine imaging at dimensions that will ultimately approach the molecular scale will depend on an improved in-depth understanding of the materials and their processing. We provide here an overview of recent work in our laboratory on the chemical and physical processes that occur during post-exposure baking (PEB) of positive-tone CA resists. Our results provide a clearer understanding of how this critical step in the lithographic imaging process will affect extendibility of the CA resist concept to nanoscale feature sizes.

Potential of DNQ/novolac and chemically amplified resists for 100 nm device generation maskmaking

An i-line optical resist based on diazonaphthoquinone (DNQ)/novolac chemistry and a chemically amplified resist (CAR) were evaluated for maskmaking application in the 100 nm device generation. By using an optimized developer with a multiple-interrupted development process, a Shipley i-line resist, SPR 700, was found to demonstrate excellent lithographic performance with a bulk sensitivity of 19 μC/cm 2 and a contrast of 10 at 50 kV. The resist can resolve 130 nm minimum feature size (MFS). The measured critical dimension (CD) uniformity of etched chrome pattern is 7.6 nm (3σ) over a field of 126 mm×126 mm. On the other hand, a CAR, Shipley UVIIHS, demonstrates a higher bulk sensitivity of 12 μC/cm 2 with the same high contrast of 10 at 50 kV as the i-line resist. The CAR also resolved 100 nm MFS and demonstrates comparable CD uniformity of etched chrome images. Both resists exhibit very good vacuum stability. The coating stability of SPR 700 is greater than 3 months, which was the length of the study; UVIIHS needs an overcoat to attain similar coating storage stability. Both resists demonstrate strong attributes for use in mask technology for the 100 nm device generation. The DNQ resist is likely to be simpler to implement than the CAR.

Simulation of surface and line-edge roughness formation in resists

A molecular-type simulation of surface (SR) and line-edge-roughness (LER) is presented and applied to a negative-tone Chemically Amplified Resist (CAR). The simulator can follow the appearance of SR and LER after each process step and predict the roughness dependence on material properties and process conditions. The simulation results are compared with SR experimental data for a negative-tone chemically amplified epoxy resist.

On-Wafer Spectrofluorometric Method for Determination of Relative Quantum Yields of Photoacid Generation in Chemically Amplified Resists

Chemically amplified resists (CARs) that employ acid catalysts are widely used throughout the semiconductor industry due to the need for high throughput in the lithography process. The quantum yield of the particular photoacid generator (PAG) used to generate a given acid ultimately limits the photospeed of the CAR. Determination of quantum yields of photoacid generation is therefore an important component of resist design. We report the development of an on-wafer spectrofluorometric technique for this purpose. This technique is based on one first reported by Feke et al. (J. Vac. Sci. Technol. 2000, B18, 136- 139), which involves doping the resist formulations containing the candidate PAGs with a fluorescent pH indicator dye, coating one wafer per PAG, patterning the wafers with a dose ramp, and spectroscopically imaging the wafers. The response curve of each PAG is spatially and spectrally encoded in the fluorescence images of each wafer. We investigate the efficacy of coumarin 6, a dye that was introduced as an acid sensor by Pohlers et al. (Chem. Mater. 1997, 9, 3222-3230) for this application. We further apply this technique to the determination of the quantum yield of photoacid generation of four candidate PAGs for prototype 193-nm CARs. This technique is convenient, fast, robust, and nondestructive.

Lithographic Performances of Non-Chemically Amplified Resist and Chemically Amplified Resist for 193nm Top Surface Imaging Process.

Application of atop-surface imaging(TSI) process by silylation to ArF lithography is desirable for ULSI production with minimum feature size for sub-120nm. We evaluated lithographic performances of non-chemically amplified positive TSI photoresist and chemically amplified negative TSI photoresist, both based on phenolic polymer. When off-axis illumination was used, 120nm and 110nm L/S patterns were obtained with non-chemically amplified resist(non-CAR) and chemically amplified resist(CAR) respectively. 100nm L/S patterns of CAR were obtained with strong off-axis illumination using binary intensity mask. CAR was superior to non-CAR in terms of lithographic performances, but inferior to non-CAR in terms of resist pattern collapse. Line edge roughness(LER) of CAR was sufficiently minimized by optimizing silylation bake temperature and it was comparable to that of single layer resist. For the prevention of resist pattern collapse in dry development process, the property of adhesion and resist rigidity is impor ant. This results can help the design of matrix resin of TSI resist for sub-100nm lithography.

Characterization of 193 nm Chemically Amplified Resist during Post Exposure Bake and Post Exposure Delay

We investigated the property change of a positive type 193 nm chemically amplified resist (CAR) during post exposure bake (PEB) and post exposure delay (PED). Upon PEB the thickness and optical property are changed. These changes depend on the exposure energy and the PEB conditions. The thickness and the imaginary refractive index changes during PEB are determined. The thickness reduction and the refractive index changes are related to the de-protection of the resist. By modeling these relationships the concentration of de-protected site of the resist and the parameters are determined. The PED effect is also considered and included in the model. The de-protection parameters needed for the 193 nm lithography are extracted. The obtained parameters are used with the lithography simulation for profile calculation, and the results are compared with the observed SEM resist profile.

Simulation of roughness in chemically amplified resists using percolation theory

A simulator for e-beam exposure and development (SELID) is combined with molecular modeling in order to investigate the various side effects of the gelation process in the case of chemically amplified resists (CARs). The procedure is demonstrated in the case of the negative tone epoxy resist but the method is equally applicable for other resist systems (positive and negative tone) and development mechanisms. A conventional resist simulator is not sufficient for the description of characteristics related to the detailed gel structure of the resist film and it needs to be combined with a molecular model. Molecular modeling is a suitable method for the simulation of the microscopic changes occurring during the post-exposure bake and development processes in the case of CARs. Macroscopic feature changes such as free volume size, cluster formation, and surface roughness can be effectively described using percolation theory. In order to construct a molecular model of a complex resist system a 3D square lattice is considered. The lattice size is equal to the mean radius of the spherical volume occupied by a monomer, in our case about 1 nm. The lattice is filled with the polymer chains by a random walk process and the photoacid generator is randomly distributed in the lattice according to its percent content in the actual material. In the case of an actual resist pattern exposure, SELID provides the energy deposition profile after e-beam exposure. The deposited energy stored into each cell of the lattice induces acid generation and subsequent reactions. Reaction progress is simulated using the molecular model. Using the above process it has been possible to reproduce the actual experimental contrast curves and to simulate line edge roughness.

Simulation of chemically amplified resist processes for 150 nm e-beam lithography

A fast simulator for e-beam lithography called SELID, is presented. For the exposure part, an analytical solution based on the Boltzmann transport equation is used instead of Monte Carlo. All-important phenomena (backscattering, generation of secondary electrons) are included in the calculation. The reaction/diffusion effects occurring during post exposure bake in the case of chemically amplified resists (CARs) are taken into account. The results obtained by the simulation are compared successfully with experimental ones for conventional and CARs. The case of substrates consisting of more than one layer is considered in depth as being of great importance in e-beam pattering. By using SELID, forecast of resist profile with considerable accuracy for a wide range of resists, substrates and energies is possible as long as the evaluation of proximity effect parameters.

Molecular dynamics simulation of gel formation and acid diffusion in negative tone chemically amplified resists

The knowledge of the structural changes that occur during the lithographic process using chemically amplified resists (CARs) is of great importance in process and resist optimization. Molecular dynamics (MD) is a suitable method for the simulation of these microscopic changes. A detailed description of the lithographic procedure including reaction propagation, acid diffusion, cage effects, free volume effects and developer selectivity can be included in such a model. The comparison of contrast curve data and acid diffusion measurements with MD modeling, leading to the evaluation of microscopic parameters are presented in this paper.

Enhancement or reduction of catalytic dissolution reaction in chemically amplified resists by substrate contaminants

Chemical interaction of resist and substrate at the interface, which modifies the dissolution reaction, has degraded sidewall profile of resist features. Depending on the nature of the residue on the substrate, the "bottom pinching" (BP) effect and footing are observed, especially for chemically amplified (CA) resists. The BP effect is observed for CA resist on top of organic bottom antireflection coating (BARC). The BP effect is attributed to the acid generated from the underlying organic BARC. With optimization on softbake temperature of BARC, the BP effect is eliminated. On a silicon nitride surface, new chemical information has been obtained which explains "footing" and BP effects in CA resists. X-ray photoelectron spectroscopy (XPS) measurements indicate that the residual alkaline molecules on the nitride surface play a major role in the formation of footing. It appears that the organic contaminants are not responsible for footing. Less severe footing is observed if the nitride surface is plasma-deposited with a thin oxide cap, which suppresses the surface basicity. However, extended plasma deposition causes resist BP. This is ascribed to the surface acidity of a newly formed oxide cap, which enhances the CA resist development process. Results show that the N (1 s) peak, after extended plasma treatment, has shifted to a higher binding state, which suggests that the nitride surface becomes acidic, causing BP.

A New High Performance CA Resist for E-beam Lithography

AbstractThree major lithographic applications have emerged for electron beam exposure tools: optical mask fabrication, direct writing for device fabrication, and more recently projection e-beam printing. The traditional mask making process uses poly(butenesulfone) resist. A wet etch process was adopted to generate patterns on chrome. Recently, shrinking dimensions, optical proximity correction features, and the complexity of phase shift masks have forced the industry to a chrome dry etch process. ZEP, a poly(methyl α-chloroacrylate-co-α-methylstyrene) based resist, has been well accepted for most of the >180 nm device mask making. The acceptance of ZEP comes in spite of its low contrast, marginal etch resistance, organic solvent development, and concerns of resist heating associated with its high dose requirements. These issues have spawned interest in using chemically amplified resist (CAR) systems for direct write and mask making applications. We have developed a high contrast resist based on ketal protecting groups, KRS-XE, which is robust against airborne contamination and can be used for all forms of e-beam exposure in both chrome mask and silicon processing. This high contrast resist is processed with aqueous base developer and has a wide bake latitude. The development of KRS-XE has provided the capability of fabricating chrome masks for future generation (< 180 nm) devices and has potential for use with projection beam exposure systems.

Application of a New Empirical Model to the Electron Beam Lithography Process with Chemically Amplified Resists

We study on the simulation models and characteristics of a highly sensitive chemically amplified resist (CAR), which is used in the fabrication of deep submicron pattern delineation. The main aim of this model is to express the positive and negative CAR by using similar principles in acid generation, reaction and diffusion, reaction mechanism, and development. Profile simulation uses the fitted parameters. Therefore, resist pattern shapes are in good agreement with the experimental ones. Simulation accuracy is 1.04% and 7.5%, for resist wall angle and critical dimension (CD) of CAR respectively. We also verify that components and properties of CAR determine the resist profile and process margin. Using this model, process feasibility of electron beam lithography is predicted, and accurate simulation is possible in the deep submicron region.

Silylation and Dry Development of Chemically Amplified Resists SAL601*, AZPN114*1, and Epoxidised Resist (EPR*1) for High Resolution Electron-Beam Lithography

A comparative study of high resolution positive imaging obtained after liquid-phase silylation and dry development with two commercial and one experimental electron-beam chemically amplified resists (CARs), namely SAL601, AZPN114, and EPR (EPoxidised Resist) is presented. 150 nm lines and spaces are obtained for all resists, while 100 nm lines are achieved with AZPN114 and EPR at 50 kV exposures. The exposure doses for AZPN114 and SAL601 are 10 µC/cm2 and 20 µC/cm2, respectively, while EPR is considerably faster (1.5 µC/cm2). Chlorosilanes are used for EPR silylation, while SAL601 and AZPN114 are silylated with hexamethyl cyclo tri silazane (HMCTS) or bis (di methyl amino) di silane (B(DMA)DS). SAL601 shows less silylation selectivity between exposed and unexposed areas possibly due to insufficient crosslinking. Implications of the presented processes on low-energy electron-beam lithography are discussed.

Characterization of acid diffusion for a negative chemically amplified resist using X-ray proximity lithography

We suggest a new method to characterize the acid diffusion phenomena for negative chemically amplified resist (CAR) using X-ray proximity lithography (XRL) in a straightforward manner. By measuring replicated critical dimensions of narrow isolated lines with various post exposure bake (PEB) times at certain PEB temperatures, we can calculate the diffusivity of CAR. This method is applied to the measurement of the diffusivity for TDUR - N908 CAR.

E-beam proximity correction for negative tone chemically amplified resists taking into account post-bake effects

The use of chemically amplified resists (CARs) in electron beam lithography required an additional processing lithography step i.e. post exposure bake (PEB). During this step the acid produced during exposure diffuses and chain reactions occur. In this paper, the method of single pixel exposures is used in order to evaluate the diffusion coefficient. This approach is applied in the case of three resists, namely AZPN114 (Hoechst), SAL601 (Shipley) and the experimental EPR. In the case of AZPN114 diffusion is limited for the higher Θ values in the center of the pattern but is considerably larger at the edge. EPR represents an extreme case where diffusion is limited by a strong cage effect due to a fast completion of the reaction while SAL601 represents an intermediate case. The diffusion coefficient is used in combination with the e-beam simulator SELID to provide data for proximity correction. The reaction-diffusion system for the specific values of the thermal processing parameters is used to modify the point spread function I. The resulting actual acid concentration for a point exposure is convoluted for a given layout. Subsequently, proximity correction is performed as in the case of conventional resists.

Critical Dimension Control in Synchrotron Radiation Lithography Using a Negative-Tone Chemical Amplification Resist

Total critical dimension (CD) controllability for 0.14 µ m line-and-space in SR lithography was evaluated for all wafer levels. The evaluation was carried out for the CD accuracy between the wafers, in the wafer and in the exposure field. The influence of the exposure process instability to the CD accuracy was also evaluated. The instability of the post exposure baking (PEB) temperature and the post exposure delay (PED) time affects the CD accuracy, and they were estimated to be less than 5 nm, respectively. The CD accuracy at the same point on the X-ray mask was 5.4 nm in the wafer and 10.5 nm between the wafers. It was found that the CD accuracy between the wafers was degraded by the inaccurate exposure dosage caused by the daily change of the SR beam distribution in the vertical direction. In the exposure field, the CD instability due to SR beam nonuniformity was 4.1 nm and that due to the X-ray mask was 15 nm. Consequently, the total CD controllability is presently estimated to be 19.5 nm for all wafer levels and the improvement of the dose repeatability and X-ray mask CD control is required to achieve the CD accuracy of less than 11 nm.