The rate of change within the microelectronics industry is not only continuing but accelerating, driven by both technical and economic requirements. Nowhere is this more visible than in the requirements for smaller and smaller geometries. 1999 will see full production of devices using 0.18 micrometers technology in Europe with 0.15 micrometers being introduced during the first years of the new Millennium; both technologies being supported by 'existing' 248nm lithography. The advent of routine subwavelength wafer patterning brings a number of major challenges for mask makers. This paper reviews the changes in mask technologies, production techniques and attitudes which will have to be adopted by maskshops over the next twelve months to meet these very real demands within Europe as the Millennium arrives.

Optimized alignment for chemical mechanical polishing has been studied in detail to ascertain the best overall overlay and process conditions. This paper describes the methodology of alignment mark design and testing in conjunction with chemical mechanical polishing optimization for technologies of 0.35micrometers and below. The planarization of the substrate material by CMP combined with asymmetric metal deposition can cause linear alignment displacement. This study investigated chemical mechanical polish slurry types, hardware configuration, and process variables on general alignment conditions. Further study on alignment mark designs and photolithography stepper settings are investigated on a subset of optimized chemical mechanical polish conditions. An alignment condition where the result of less than 0.10 micrometers was obtained.

High voltage bipolar and BiCMOS processes often use thick epitaxially grown layers of silicon. These films 12-24 micrometers thick offer a considerable challenge to the alignment of subsequent process layers due to the 'wash out' and image distortion, caused to any underlying pattern, which render automatic alignment mark recognition difficult it not impossible. Historically using projection aligner technology these immediately post Epi layers have been manually aligned with future automatic alignment target defined at the first opportunity post Epi. This is not possible using ASM steppers, as these depend upon marks etched into the silicon, before first processing, to create marks, to which all subsequent layers are registered. To allow the stepper to run wafers with these Epi films a new approach was required.

In this paper, the current status and recent progress in the field of deep UV lithography is reviewed. The introduction of resolution enhancement techniques and high NA 248nm lenses is discussed, with emphasis on their impact on intrafield linewidth control. It is expected that 248nm will be used for volume manufacturing of the 0.15 micrometers gate length devices and perhaps even pushed to the 0.13 micrometers generation. The current status of 193 nm lithography is reviewed, which is expected to be inserted at the 0.13 micrometers technology node. Based on the current situation at 248nm, the extendibility of 193nm lithography to the 100nm node is discussed.

Lines and spaces with posts are a typical combination of feature types for metal layer applications. A common process window for these is difficult to achieve, especially for equally sized features on the wafer. Foremost in attaining one is minimizing the differences in the dose to size of the feature. Many methods have been studied to maximize overlapping process windows of different features. Direct feature biasing, assist features and running NA and sigma are all methods that can be applied to maximize overlap. In this work, two different feature types, isolated and dense lines and posts, are examined. It is shown, through simulation and experimentation, how the overlapping process window (OLPW) of these features can be optimized. PROLITH software will be used to simulate OLPW with the latest resist models. This is examined experimentally, using an ASM 5500/300 DUV stepper with variable NA and partial coherence, for 250nm dense and isolated lines and 250 nm posts at optimal illumination settings which are determined by PROLITH. The improvements in OLPW due to resist type, e.g.; positive tine UV6 and negative tone UVN30, combined with optimal illumination coherence and mask bias, are shown. Mask feature bias is examined for the amounts that produce a common process at each of three partial coherence settings. Also given are the trends for isolated to dense and line to post proximity bias and the comparison of these to simulation.

For many generations of integrated circuit technology optical lithography has been used for patterning. Given this dominance the possible limits of optical lithography are examined and the challenges towards reaching these limits are set out. The present and near future technologies are investigated in detail and the more long term technologies are looked into and finally a possible optical lithography exposure system roadmap is constructed.

Photolithography has now moved forward to such an extent that we are considering imaging the 130nm and 100nm technology nodes with optical system, unthinkable a handful of years ago. To do this we have been using the biggest control knob we know, wavelength. The latest wavelength to be introduced into production will be 193nm which will arrive with full field scanners this year. To image using 193nm we have had to radically change the polymer systems we use due to high absorption with conventional chemistries. This has led to materials that will be the most difficult to integrate into manufacturing that we have ever faced. The primary challenge we encounter is the high resist thickness loss in the pattern transfer steps. This is primarily due to low etch resistance of the materials in use but we will also show that photochemical deprotection of the resist during etch has a contribution. One approach to overcome this is to use significantly thicker resist films, this leads to its own problems with pattern collapse a major worry which could easily become a limiting factor in this technique.

The ALTA 3500, an advanced scanned-laser mask lithography tool produced by Etec, was introduced to the marketplace in 1997. The system architecture was described and an initial performance evaluation was presented. This system, based on the ALTA 3000 system, uses a new 33.3X, 0.8 NA final reduction lens to reduce the spot size to 0.27 micrometers FWHM, thereby affording improved resolution and pattern acuity on the mask. An anisotropic chrome etch process was developed and introduced along with a TOK iP3600 resist to take advantage of the improved resolution. In this paper we will more extensively describe the performance of the ALTA 3500 scanned laser system and the performance of these new processes. In addition, an investigation of the benefits of operating in the optimal isofocal print region is examined and compared to printing at the nominal process conditions.

This work discusses routes to extend optical lithography to the 70 nm technology node using proper selection of masks, mask design including choice of optical proximity correction (OPC), exposure tool, illuminator design, and resists design to do imaging process integration. The goal of this integration is to make each component of the imaging system work to the best benefit of the other imaging components so as to produce focus-exposure process windows large enough to use in a manufacturing environment.

OAI and PSM which are effective to improve the resolution and DOF are new accepted as key technologies by extending the limit of optical lithography. However, for these techniques, it is very difficult to control the pattern size error within 10 percent as the pattern size becomes smaller. The physical mechanisms of resolution and DOF enhancement of the new annular illumination and the new mask is briefly discussed. Some result of simulation to investigate the basic lithographic characteristics of the new annular illumination and the new mask are presented. The structure of the mask is presented in this paper. Finally, experiments have been carried out to verify the characteristic of the new annular illumination and the new mask, and the results verified the enhancement of lithographic performance. In i- line exposure system, we obtained best resolution of 0.35 micron and 2.4 micron DOF for 0.6 micron feature size. In our simulation, with 193 nm exposure wavelength, 0.13 micron pattern size can realized with 2.0 micron DOF, and the proximity effect decreased obviously.

The implementation of new resist materials and advanced lithography processes requires new characterization techniques in order to understand the behavior of these systems and optimize their design. Examples are presented on two recent 193nm experimental formulations using three different characterization techniques, namely Modulated- Temperature DSC, in-situ ellipsometry and dielectric analysis. The results obtained provide new experimental evidence of the diffusion and reaction mechanisms involved in chemically amplified formulations.

In this paper a new method to determine photoresist Dill parameters is presented. Based on spectroscopic ellipsometry (SE) measurements, this new method is more precise than standard techniques based on transmittance measurements. Indeed, compared to photometry, SE technique is a self calibrated technique which provide directly two independent parameters Tan (Psi) and Cos (Delta) which can be used to extract directly thickness but also optical indices of a layer inside a multilayer structure. Moreover, the wavelength dependence introduces more restrictions for the data analysis since thickness and optical indices can be deduced directly in many cases. We apply this technique to different kinds of photoresist designed for 365nm and 248nm. At each wavelength ellipsometric parameters are simulated directly versus the exposure dose without any assumption on the thickness and on the index of refraction evolution. On 365nm photoresist this new method provides Dill parameters in good agreement with the standard method. On 248nm photoresist we show that the influence of the exposure is more important on the refractive index and on the thickness of the layer than on its absorption.

Spectroscopic ellipsometry has long been recognized as a powerful technique for nondestructive characterization of thin films and multilayer structures. In the case of photolithography it has been intensively used for layer thickness and optical index determinations. Nevertheless, the main drawback of this technique is that it is not direct. The data analysis must be made adjusting a prior optical models on the experimental data. For photolithographic films, it is sometimes difficult to find accurate structural model and the analysis becomes difficult when the film is absorbent in the entire wavelength range. It is why SOPRA has developed recently a new versatile instrument that integrates different experimental techniques. The base is the SOPRA GESP5 instrument which offers already the possibility to perform spectroscopic ellipsometry, spectroscopic photometry and scatterometry. We have added a grazing x-ray reflectometer option, which allows measurements on the same spot at the same time, and combine analysis of the different data. The source is a compact ceramic Cu fine focus x-ray tube. the x-ray beam is defined by Soller Slits and a parallel interchangeable slits. After reflection on the sample surface, the beam is monochromatized using a curve graphite crystal and focused on the detector. This new system is discussed in details hereafter and characterization of antireflective for microlithography is presented.

Among the chemistries/polymers reported for the 193nm photoresist applications, methacrylate copolymers consisting of 2-methyl-2-adamantane methacrylate and mevalonic lactone methacrylate and cycloolefin polymers derived from derivatives of norbornene have shown promising results. We have studied the lithographic properties of these two but different promising chemistries. Both system offer linear resolutions down to 0.13 micrometers using conventional 193 nm illumination and high sensitivity at standard developer conditions. While the methacrylate based system shows best performance on substrates with bottom coats, the cycloolefin-Maleic anhydride alternate copolymer based resists performs well on bare silicon as well as substrates with bottom coats. The etch rates of the methacrylate and cycloolefin based resists were found to be 1.4 and 1.3 times relative to that of KrF resist APEX-E. Further, new polymers consisting of isobornyl and alkyl ether chains on the ester groups of norbornene carboxylate were made in order to decrease the glass-transition temperatures of the norbornene-maleic anhydride type polymers. These results will be included and discussed in detail.

Leading-edge technologies require continually shrinking design grids due to the industry demands for decreasing minimum feature size and higher resolution. Using conventional raster-scanned exposure tools to place these patterns on photomasks results in longer writing times, because linear decreases in address result in exponential increases in writing time. This phenomenon can be compensated throughput at small addresses while retaining lithographic quality.

A scanning electron microscope (SEM) has been modified for direct-write electron beam lithography. The instrument has the capability to automatically align with the features patterned by optical lithography and exposure the features requiring the finest linewidth with the electron beam. The main application for the instrument is one the process line for fabricating high electron mobility transistors (HEMTs) in GaAs monolithic microwave integrated circuits (MMICs). The high frequency performance of the HEMTs is critically dependent on the length of the gate electrode and the placement of this electrode between the source and drain electrodes. All of the mask layers for a MMIC except the gate layer are exposed using optical lithography, as it provides the required linewidth with high throughput. The sample containing the partially fabricated HEMTs is loaded in the instrument and is positioned at each HEMT in sequence in order to acquire an image of the source and drain electrodes. This image is correlated with a reference image of the HEMT to determine its precise location for subsequent exposure of the gate electrode by the electron beam. The instrument is able to achieve an alignment accuracy of 80 nm and has been used to expose features with linewidths less than 100 nm. As images of the device are used for alignment, the instrument does not require alignment marks on the sample and is able to automatically compensate for positional errors caused by same stage and mask tolerances. As the full SEM functionality of the instrument is retained, it may also be used to inspect the results of the lithography.

This paper describes work carried out to evaluate the performance of Shipley UV5 DUV positive tone photoresist for electron beam lithography using experimental design techniques. Factors which affect the results of resist processing have been identified by initially performing two level factorial screening test followed by detailed response surface analyses. Lines and spaces with a 120 nm period were produced using optimized conditions and there was no evidence of process delay effects over three hour periods in air. Separate batches of the resist were found to have similar optimum baking temperatures. The result obtained have been used to develop reliable techniques for the fabrication of 'T' shaped and Gamma shaped metallized gates for high frequency circuit applications. Comparisons have been made with result obtained using Shipley UV3 photoresist. We conclude that UV5 resist is a useful resist for electron beam lithography applications.

The effects of developer temperature on several conventional resist and one chemically amplified resist, and the effects of developer normality on the dissolution behavior of a 248nm chemically amplified resist, are examined using development rate measurements. Using an RDA-790 development rate measurement tool, dissolution rates as a function of dose and depth into the resist were measured. Each data set was analyzed and the performance of rate versus dissolution inhibitor concentration was fit to appropriate models. The variation of these results with developer temperature has led to temperature-dependent characterization of the dissolution modeling parameters. The variation of dissolution rate with developer normality has led to an initial characterization of the normality-dependent dissolution modeling parameters.

In general, simulation requires a thorough understanding of the physics and/or chemistry of the processes. This should lend itself of models which can be used to establish simulation software. In addition, for a simulation to be successful, a calibration of the model is needed. A good model using bad parameters returns bad results. In lithography simulation there are settings of parameters which are well known. Others are less known and may be hard to obtain. A typical example is the development parameters, or parameters describing the reaction mechanism for chemically amplified resist. To support the user of simulation software in the process of finding proper input parameters, the new software package FIRM has been developed and will be presented in this paper, together with applications. FIRM uses models for the optical or e-beam lithography, the same as SOLID-C and SELID, and determines any set of coefficients from given experimental observations. From an initial set of coefficients, it tries to fit calculations to observations. FIRM accepts various types of measurements, e.g. thickness tables of the resist or focus-exposure matrices. In addition, the user selects from a wide list of resist models the parameters to be refined. FIRM then tries to find correlation between the parameters and the differences between calculation and observation. In an iterative process 'best' parameters are determined. The validity of the algorithm is verified against well known test cases. Next, applications of FIRM to several new chemically amplified resists for DUV will be presented using different types of experimental input.

TCAD simulation is very important for DUV lithography process development and control. Traditional lithography process engineering has relied on short-loop and pilot-lot experiments to understand the effects of particular process control factors. However, experiments are very expensive, and the complexity of lithographic patterns and processes is such that we must often resort to computational simulation. The availability, accuracy, and ease of use of lithography simulation are essential to the semiconductor industry. In this paper we present a methodology for DUV lithography simulator tuning by resists profile matching. A global optimization procedure is used to efficiently extract the correct values of the important fitting parameters by matching the simulated resist profiles to measured data. Results for a DUV lithography process are presented.

There are shown the basic experimental result, achieved by the 1.06-micrometer intracavity laser processing of the 800 angstrom aluminum film evaporated on the optical glass substrate. There is described the nw laser cavity geometry, which allows to form 2D submicron-size image on the solid- state surface by means of one pulse of laser radiation.

An experimental and theoretical study of a coherent multiple imaging technique that utilizes a Fabry-Perot etalon placed between the photo mask and the projection lens is reported. This technique can enhance both resolution and depth of focus in optical microlithography. A lithography simulation tool, Prolith/2 was used to evaluate the aerial image profile using a complex phase-amplitude pupil-plane filter to simulate the effect of the Fabry-Perot etalon. This work specifically discusses the evaluation of extended periodic patterns, widely used in lithographic simulations. Simulation results are described and compared with experimental data. The impact of Talbot images generated by periodic structures is also described.

The best focus shift problem was observed on PAS-5000/50 I- line system. The best focus shift could induce poor resist profile in back-end lithography exposure processes, and caused blind via-holes or damage on conductor lines after plasma etch. The effects of topography change, pattern layout and focus laser spot position on the best focus determination were proved related to the focusing system design, single point focus detection with small size probe beam. of PAS-5000/50 steppers. In the main paper, the weak points of the design are pointed out, i.e. the unwanted information due to its dependence on step topography and layout may be collected and cause defocused resist profile during focusing. Though the application of dark-tone digitized scribe line design on dielectric layers reduces step topography discrepancy between main cell and scribe line and partially solves the problems; however, the complete solutions rely on the fundamental improvements on the stepper focusing algorithm.

Critical Dimensions (CD) are measured on reticles which are written incorporating an optical proximity correction for pitch dependent linewidth bias. Reticles are manufactured on different generations of laser write tools and the result in terms of mean ann range of dimensions compared. Lithography simulation is used to test the response of reticle CD variation at wafer level.

An analysis technique for quantifying the effectiveness of optical proximity corrected (OPC) photomasks is described. The methodology is able to account for reticle manufacturing tolerances and has a number of applications including the optimization of OPC features and, in the examples described, the analysis of defect printability. The results presented here are generated using aerial image measurements from PROLITH/2, but the technique can be directly transferred to resist image measurements using 3D simulation tools such as PROLITH/3D where other factors such as swing curve effects caused by wafer topography could also be analyzed. With inspection tools such as scanning electron or atomic force microscopes and appropriate image processing and analysis software it should also be possible to apply this methodology to practical results.

Resist contrast is related to usable depth-of-focus (UDOF) for I-line and DUV stepper lenses through the resists process linearity function. Results, supported by simulations in the resist image, show that stepper field center-edge critical dimensional offsets, which decrease UDOF, increase as resist contrast decreases. Measurements on intra-field CDs show that their uniformity decreases as processes reach non-linear processing regimes.