X-ray Stepper Research Articles

Application of interferometric broadband imaging alignment on an experimental x-ray stepper

A series of experiments were performed with an interferometric-broadband imaging (IBBI) alignment system on an experimental x-ray lithography stepper. These experiments demonstrated sub-1 nm consistency of independent IBBI measurements and the ability to feedback lock the mask relative to a wafer to within a mean of 0.0 nm and a standard deviation of 1.4 nm. Comparisons of displacement measurements made with IBBI and closed-loop piezo drives confirmed scale consistency to within about 1.5%. In the absence of feedback control, spurious relative displacements were observed, due to temperature gradients and the nonrigid mechanics. This argues in favor of feedback control during exposure. The robustness of IBBI in allowing nanometer-level alignment measurements during x-ray exposure, with remotely located long-working-distance, low-cost microscopes should enable it to be used as a stand-alone alignment system or as an adjunct to global alignment.

Alignment accuracy evaluation of the X-ray stepper SS-1 for processed wafers

The alignment accuracy of the X-ray stepper SS-1 for synchrotron radiation (SR) lithography was evaluated. The mark state changes whenever the process layer is deposited. The alignment mark can be classified into four types based on the mark composition of all the process layers in LSI manufacturing. The alignment accuracy was evaluated by these four typical process marks. For a silicon mark, alignment repeatability of 25 nm (3σ) was obtained. For the renewal mark composed of SiO2 or metal layer, alignment repeatability of 60 nm (3σ) or less was obtained. In all the process layers used in the experiment, alignment repeatability of 80 nm (3σ) or less was obtained. The signal level can be improved by optimizing the step depth of the alignment mark and by optimizing the process layer thickness. The same alignment accuracy as that of a silicon mark is possible by optimizing the mark step depth and the process layer thickness.

A novel X-ray mask concept for mex&match lithography fabrication of MOS devices by synchrotron radiation lithography

The investigated X-ray steppers Karl Suss XRS 200/1 and XRS 200/3 use an optical alignment system for adjusting the parallelism, for gap setting and alignment between mask and wafer. In general the optical signal, when mask and wafer are at gap distance, is dependent on the optical characteristics (transparency of the mask, reflectivity of the wafer, resist film interference etc.) of both mask and wafer. In this paper, different methods to decouple the optical signal of the mask by that of the wafer and to effectively improve the optical signal contrast, are presented. The methods used allow to reach in all examined cases an alignment accuracy down to 5 nm. This value can be regarded as the resolution limit of the alignment system itself.

Dynamic motion of mask membrane in x-ray stepper

This article presents a new analysis method and experimental results for the deflection and vibration of x-ray masks, which are caused by the squeeze effect of the gas film between a mask and a wafer in x-ray steppers. The deflection and vibration occur when a mask-to-wafer gap setting is executed or if the wafer vibrates in the mask direction during stepping motion with a narrow gap. They are not only detrimental to the throughput, but may also damage the mask membrane. Lees’ difference approximation method is applied to a compressible Reynolds’ equation and the equation of motion for the mask membrane. The main results of calculations and experiments, which showed good agreement with each other, are outlined below. (1) When the wafer approaches the mask at a constant velocity, the mask deflection increases in proportion to the velocity, and the mask deflection is larger for a smaller mask tensile stress. (2) When the wafer vibrates in the mask direction, the amplitude of the mask vibration increases with increasing the wafer frequency, and reaches a maximum value at a frequency that depends on the gap size. The maximum amplitude of a tested mask becomes 4 times as large as the wafer amplitude. (3) For a high frequency and narrow gap, the mask vibrates while deflecting convexly in the direction opposite to that of the wafer.

X-ray induced mask contamination and particulate monitoring in x-ray steppers

The belief that proximity x-ray lithography is insensitive to contamination has been based on the assumption that contaminants are small hydrocarbon compounds. However, we will show that x rays may induce photoassisted processes that deposit saltlike films on the mask surface. We will also provide an explanation for the film geometry in the early stages of its growth and suggest some possible solutions to the issue.

Installation and initial operation of the Suss Advanced Lithography Model 4 X-ray Stepper (abstract)

A Suss Advanced Lithography X-ray Stepper designed as a production tool for high throughput in the sub-quarter-micron device range has been installed and is being commissioned at the University of Wisconsin’s Center for X-ray Lithography (CXrL). Illumination for the stepper is provided by a scanning beamline designed and constructed at CXrL. The beamline optical components are a gold-coated plane mirror, a 1-micron-thick silicon carbide window, and a 25-micron-thick beryllium exit window. Beamline features include synchronized scanning of the mirror and exit window, variable scan velocity to compensate for reflectivity changes as a function of incident angle, and a horizontal oscillation of the beryllium window during vertical scanning to average the effects of nonuniform beryllium window transmission. A helium purged snout transports the x-rays from the beamline exit window, to the exposure plane in the stepper. This snout is retractable to allow for the loading and unloading of masks into the stepper. The motions of the mirror, exit window, and snout are computer controlled by a LABVIEW program that communicates with the stepper control software. The design of the beamline and initial operating experiences with the beamline and stepper will be discussed.

An overview of X-ray lithography

Current achievements and future prospects of SR base x-ray lithography are explained through an account of research and development efforts made by NTT LSI Laboratories. The SR lithography system, including the compact SR light source Super-ALIS and beamlines with efficient mirrors, is now at the stage of practical use. The vertical x-ray steppers SS-1 and SS-2 have attained a repeatability of overlay accuracy as high as 25 nm (3σ). High-accuracy x-ray masks can be fabricated by applying the multiple-exposure method. Further improvement in pattern placement accuracy is expected with application of distortion compensation writing. Intense efforts made in development of x-ray mask inspection and repair technology resulted in the fabrication of defect-free x-ray masks. Test fabrication of CMOS/SIMOX LSIs has yielded fully functional devices, with an improved overlay accuracy of 0.1 to 0.15 μm (3σ), and a CD control of ±10% in 0.2 μm regions. Further application of SR lithography to the test fabrication of future LSIs will prompt the implementation of this technology into the practical process.

Sub-0.1 μm NMOS transistors fabricated using laser-plasma point-source X-ray lithography

We report the experimental results of the first MOSFET's ever fabricated using a laser plasma-source X-ray stepper. The minimum gate length of these transistors is 0.12 μm with an effective channel length of 0.075 μm. These transistors were patterned using a mix-and-match lithography scheme where the gate level was printed using a 1.4 nm plasma-source X-ray stepper while the other layers were patterned using optical lithography.

Lithography for micro-electronics

Synchrotron-radiation-based X-ray lithography has been studied for more than 15 years as an alternative to conventional UV lithography. Storage rings, beamlines, X-ray masks, X-ray steppers and other components have been developed and optimized for X-ray lithography. Compact rings have therefore become commercially available, beamlines have been optimized for practical use, and test devices can now be fabricated because of improved total accuracy in the X-ray masks and steppers. This report discusses the status of key techniques involving synchrotron radiation sources, beamlines, X-ray masks, and X-ray steppers.

A unified approach to resists materials design for the advanced lithographic technologies

New resist materials and processes are necessary to pattern ≤ 0.25 μm design rule circuits with advanced deep-UV, X-ray and e-beam lithographic technologies. Chemically amplified positive resist systems introduced to meet the high sensitivity and resolution requirements of the deep-uv, e-beam and x-ray exposure tools suffered from marginal adhesion, poor etch resistance and deteriorating process performance with post-exposure delay (PED) time. Improved resist systems such as those based on materials poly(4-acetoxystyrene-4-t-butoxycarbonyloxystyrene-sulfone) (PASTBSS) terpolymers resolved the adhesion problems and improved the etch resistance and post-exposure delay time stability. Theses resists, however, still required a covercoat for good process performance. Additionally, all the current commercial chemically amplified positive resists show varied degrees of “foot” formation on Titanium Nitride and Silicon Nitride substrates and strong linewidth dependence on PEB temperature (large Δlw/°C). We have developed a new multi-component positive chemically amplified resist called ARCH (Advanced Resist CHemically Amplified), that in addition to exhibiting excellent resolution with deep-uv, x-ray and e-beam exposures, displays no noticeable foot on Titanium Nitride, Silicon Nitride and BPSG substrates. Initial results with deep-uv lithography also indicate that linewidth dependence on post-exposure bake (PEB) is minimal. This chemically amplified resist system exhibits linear 0.25 μm resolution with a GCA XLS excimer laser stepper (0.53 NA, 248 nm) and 0.14 μm resolution with a pulsed laser point source proximity print x-ray stepper (λ centered at 1.4 nm) and 0.1 μm resolution with a JEOL JBX-5DII e-beam exposure system (at 50 keV).

Performance of the Modified Suss XRS 200/2M X-Ray Stepper at CXrL

An XRS 200 model 2M X-ray stepper has been installed at the Center for X-ray Lithography. This paper describes the capability of the machine and its installation. Of greater interest is the performance the stepper gives with respect to the dose control and the mask/wafer alignment. How robust a process is depends on the latitude of each process parameter and the control tolerance accorded the machines and materials that affect each one. The dose repeatabillty and uniformity is determined by the optical elements in the beamline, the helium delivery system, and the accuracy of the scanning stage. Test results are reported for dose control using three separate methods. The scan speed is measured using Suss supplied equipment and is verified separately by CXrL's own technique. The dose delivered is analysed by the response of radiachromic film. Also important is the behavior and performance of the ALX-100 alignment system. The optical set-up is described as a model that includes the ALX hardware and illumination as well as the mask and wafer. The signal and contrast from the mask and wafer are compared to data for several substrates. A figure of merit is discussed for determining which membranes will give the best overlay.

High-performance multilevel blazed x-ray microscopy Fresnel zone plates: Fabricated using x-ray lithography

Diffractive lenses are becoming the optical elements of choice for many applications. One type of diffractive lens, the binary zone plate, has already demonstrated high-resolution performance experimentally. However, in order to increase the diffraction efficiency of these zone plates, a blazed grating profile must be used. This can best be approximated by a staircase grating profile, created by multilevel exposures. Using x-ray lithograph, we fabricated for the first time circular, linear bi- and trilevel zone plates, with gold structures 0.75 μm thick (per level), on silicon nitride substrates. The zone plates were designed for use at a wavelength of 1.54 Å, and had a theoretical efficiency of 68.5% for bilevel and 81.5% for trilevel zone plates. Due to the large depth of focus and high resolution inherent to x-ray lithography, the finished zone plate exhibits very steep sidewall profiles, with linewidth resolution down to 0.25 μm. Such vertical sidewalls are essential for achieving high lens efficiency. Fabrication errors, such as thickness variation in the electroplated gold and misalignment, were considered, and their effect on the optical efficiency of the zone plate was estimated. Alignment errors between levels were minimized, achieving a best result of 25 nm (3σ). In fabricating the zone plates, we employed standard integrated device tools, such as a Leica Cambridge Electron Beam microfabricator (EBMF) 10cs/120 electron-beam writer for the x-ray mask fabrication, and a Suss 200 x-ray stepper for the multilevel exposures. Thus, we have shown that it will be possible to fabricate many lenses, with a variety of optical characteristics, in one wafer.

Applicability test for synchrotron radiation x-ray lithography in 64-Mb dynamic random access memory fabrication processes

To evaluate the applicability of synchrotron radiation x-ray lithography (SRXL) in ultra-large-scale integration manufacturing processes, we simulate a part of a dynamic random access memory process using SRXL. Four levels of x-ray masks (field, gate, bit contact, and bit line), including fiducial patterns for coordinate/overlay measurement, alignment marks, and 0.4-μm-rule memory cells were fabricated using an i-line stepper. The maximum overlay error between the different level masks was 60–100 nm. In SRXL processes alignment errors of the x-ray stepper were 60–100 nm (3σ), and the overlay errors over a 20-mm-exposure field were 130–160 nm (‖mean‖+3σ) in most of alignment levels. Critical dimension control of 27–47 nm was obtained for gate length and contact hole diameter.

Evaluation of overlay accuracy for the x-ray stepper TOXS-1

This article presents the overlay accuracy for the newly developed prototype x-ray stepper TOXS-1. Checkerboard gratings on a mask and wafer were used for an optical-heterodyne interferometry alignment system so that the alignment signals from the mask and wafer gratings can be detected independently without mutual interference. The alignment signal varied slightly with the mask-to-wafer gap due to multiple reflection of the alignment beams between the mask and wafer, which deteriorates the alignment accuracy. To reduce the multiple reflection, the mask grating and the mask alignment window were coated with opaque film and antireflecting film, respectively. A 0.025 μm (mean+3σ) overlay accuracy has been achieved for SiO2 processed wafers. The overlay accuracy was further measured for four different kinds of processed wafers, and a 0.035 μm (mean+3σ) accuracy has been obtained except for an aluminum processed wafer.

Impact of chuck flatness on wafer distortion and stepper overlay - comparison of experimental and FEM-results -

Abstract Overlay accuracy is one of the most important lithographic topics for ULSI device production [1]. Significant contributions to the overlay like alignment accuracy and mask distortion are well known and subject of continuous investigations [2]. Below an accuracy range of 100 nm additional influences must be taken into account [3,4], which were usually neglected for relaxed design rules. One of these influences is directly related to wafer distortions induced by flatness inaccuracies of wafer chucks used for instance in X-ray or optical stepper. This impact was characterised by investigating the elastic behaviour of 4” wafers, fixed on a wafer chuck. The resulting strains and elongations in the wafer surface were measured by exposure experiments and compared with additional FEM calculations. From these results we estimate that for design rules below quarter micron the flatness accuracy between different wafer chucks has to be smaller than 0.25 μm.

Replication of very small periodic gratings with proximity x-ray lithography

Abstract This paper describes our recent results in high resolution synchrotron radiation lithography. Theoretically we show an image analysis based on the absorbed dose distribution in the x-ray resist rather than on the diffraction intensity calculations. Experimental results are presented, demonstrating a successful replication of very small periodic gratings using high resolution SiC/W masks and a high performance commercial x-ray stepper.

A new optical heterodyne alignment and gap detection technique for x-ray steppers

A new optical heterodyne interferometry alignment technique, which measures the displacement and gap between mask and wafer has been proposed. Both the lateral displacement and the proximity gap can be detected using the same optics with two pairs of laser beams with slightly different frequencies. Optical symmetry is essential in this technique. The principle of operation, the optical system and the experimental results are described. The capability for detection of displacement and gap of as small as 11nm and 0.37μm, respectively, was experimentally confirmed.

ＳＲ転写用ビームラインシステム

A high power X-ray delivery system consisting of a superconducting compact storage ring and a novel beamline is now in use for SR (synchrotron radiation) lithography research at Mitsubishi Electric. The double ellipsoidal focusing mirror optics of the beamline collects and condenses the horizontally diverging SR beam and transports it to the exposure station, a Karl Suss X-ray stepper. The beamline is kept at an ultrahigh vaccum level by the vacuum control system. Furthermore, the vacuum protection of the beamline ensures that a storage ring vacuum of at least 1.33×10-6 Pa is maintained, even when the beamline vacuum is broken down. We have also investigated the effects of the electron beam orbit change on the vacuum performance of the beamline.

Improvement of heterodyne alignment technique for x-ray steppers

This article describes the alignment accuracy of the x-ray stepper that was codeveloped by the SORTEC Corporation and Matsushita Electric Industrial Co., and installed in the Tsukuba Research Laboratory for the SORTEC Corporation [K. Koga et al., J. Vac. Sci. Technol. B 8, 1633 (1990)]. Although the optical heterodyne method used for the stepper has a high position detection sensitivity, this method has the drawback that the alignment accuracy is apt to be influenced by an uncontrollable interference, because it uses a coherent laser light source. The overlay accuracy of the stepper was evaluated and identified the error factor contributing to the total overlay error. It was found that the deterioration of the alignment accuracy resulting from position detection error is caused by subtle deflection in the optical alignment system. In order to solve this problem, the alignment unit of the x-ray stepper has been improved. As a result, an alignment accuracy of less than 24 nm (3σ) was obtained in a double-exposure experiment with a posiresist. In addition, an alignment accuracy of less than 38 nm (3σ) was also obtained with SiO2 etched marks.

Overlay performance of x-ray steppers in IBM Advance Lithography Facility

This article describes the overlay performance of the Suss XRS200/3 synchrotron compatible x-ray steppers installed in the IBM Advanced Lithography Facility. The Suss XRS200/3 uses a video-based brightfield microscope alignment system. The overall performance of the stepper was described previously. This article will specifically focus on the test, optimization, and performance of the alignment system. In the Suss alignment system, the quality of optics and test methods of the optical assembly was a critically important factor in achieving the 110 nm overlay quoted previously. A detailed discussion is given on the selection and characterization of these alignment optics, along with the subsystem level tests used to ensure performance of the complete imaging assembly prior to and after installation on the main stepper body. The overlay test methodology and data obtained during the acceptance testing is also discussed in detail. The exposure testing was designed to measure a full range of performance of the stepper. Indeed, overlay ranged from 50 nm (mean plus three sigma) on the very best levels to in excess of 150 nm on the most difficult. Sensitivity of the alignment system to small film thickness changes was tested along with alignment precision as a function of alignment mark depth or signal contrast.