摘要:

The placement of pattern features in lithography must be accurate to within some small fraction of the minimum feature size or critical dimension (CD). As CDs decrease over the years, driven by the demand for high density memory, the requirements on pattern placement become accordingly more stringent. A practical problem exists: after a wafer or mask is written, a binary decision must be made, based on a set of measurements, whether the pattern placement is deemed acceptably accurate. Based on this decision, the substrate will either be retained for further processing, or discarded. The large number of pattern features, together with the slowness of present‐day metrology, cause the acceptance decision to be based on measurement of a tiny fraction of all features in the pattern. It is a classic problem in sampling statistics. A proper analysis relies on the probability thatanyfeature, were it to be measured, would have a placement error exceeding some predetermined limit. In practice errors contain both systematic and random components. Systematic errors include magnification, orthogonality, and distortion, to name a few. Random errors arise from noise and jitter in the lithography or metrology tool, for example. The purpose of this paper is to derive a criterion for the binary acceptance decision, which is useful in the presence of combined systematic and random errors. The distribution of systematic errors is shown to be proportional to the area between the lines of a contour map of the error. The method is applied to an e‐beam written mask for x‐ray lithography. It is shown that conventional variance analysis is inadequate for predicting the error rate.

ISSN:0734-211X    

DOI:10.1116/1.585844

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

The atomic force microscope (AFM) has the potential for profiling on a nanometer scale. However, for the AFM to be useful as a reliable and accurate metrology tool, the atomic force sensor used (i.e., cantilever and sensor tip) should be highly reproducible in the fabrication process and the tip shape used should be known with high accuracy. Moreover, for profiling and critical dimension measurement of device trenches with submicron openings and trench depth of 1 μm or more, nanometer‐scale tip shapes with high aspect ratio will be required. In order to meet the above requirement, a novel high‐resolution direct electron‐beam deposition process has been developed in which a sharp tip is grown onto an existing Si cantilever structure. In the example to be discussed, sharp tips have been grown directly on the integrated tip of a micromachined Si cantilever by decomposition of a dimethyl‐gold organometallic complex. The directe‐beam deposition process provides a unique capability for the controllable fabrication of high aspect ratio, nanometer‐scale tip structures. Typical dimensions of the AFM tips achieved are 0.1 μm in column diameter, 1–4 μm high, and 15 nm in tip radius. Recent feasibility experiments of using thesee‐beam tips in the laser force microscope have demonstrated for the first time that this nano‐scale tip can be used for mapping 1‐μm‐deep Si trenches. Submicron trenches with openings as small as 0.36 μm have been successfully profiled. In this paper, various issues on tip fabrication, deep trench profiling and comparision of AFM and corresponding scanning electron microscopy measurements will be discussed.

ISSN:0734-211X    

DOI:10.1116/1.585845

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

We have developed a novel technique for producing tips for scanning probe microscopy. The shape of the tip is optimized for applications where high aspect ratio surface topography is the norm, as with integrated circuit structures. The technique involves focused ion beam (FIB) milling of a tip which was previously shaped to a nominal geometry by electrochemical etching. The ion milling pattern is annular, and the ion beam is collinear with the axis of the tip. The process allows control of the ion milling dose using 20 keV Ga+with submicron control of the annular patterns. The result is a narrow, tapered structure approximately 20 μm in length which ends in a point with a radius of curvature between 30 and 50 nm when grains dominate the sputter process, and radii of about 3–4 nm when there is no evidence of grain structure effects. This microstructure is 3 μm in diameter at the base and it protrudes from a portion of the shank of the macrostructure where the diameter is about 15 μm. We have sufficient control over the sputter process to yield the final tip length, taper, and radius. Cone angles between 12° and 15° over the first two microns from the apex can be achieved routinely, by the correct choice of annulus and ion dose. Sputter simulations predict the correct shape of the final tip profile, and show the effect of varying the ion beam focus, dose, and inner and outer annulus radii. Tips with the desired geometry have been produced in polycrystalline tungsten, iridium, and platinum‐iridium. Significant improvements in scanning tip microscopy (STM) images have been consistently observed with these FIB milled tips.

ISSN:0734-211X    

DOI:10.1116/1.585846

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

Critical dimension metrology relies on the analysis of electron‐beam signal profiles from structures of arbitrary size and shape. Conventional modeling techniques for simulating such profiles use Monte Carlo techniques which, while very accurate, require considerable computing time since any change in any parameter necessitates a complete new calculation. This paper describes an alternative approach in which a Monte Carlo simulation is used to derive the coefficients of a diffusion matrix. From this matrix the backscattering and secondary electron profiles from a surface of given geometry can be found by simple summation, and the results of a change in the geometry can be determined immediately without the need for further computation. The predictions of this method are in excellent agreement with those of the conventional approach.

ISSN:0734-211X    

DOI:10.1116/1.585847

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

The form of the screened Rutherford elastic scattering cross section has been modified to approximate the Mott scattering cross section for low energy electrons interacting with high atomic number targets, both for the total and the differential elastic cross section. A modified form of the total scattering cross section has been found that fits the Mott scattering cross section derived using a partial wave expansion. The fit is from atomic number 6–92 over the energy range 1–100 keV. The differential elastic cross section has been modeled using a screened Rutherford distribution plus an isotropic distribution. The ratio of forward to backscattering found from the Mott cross section was used as a fitting criteria. The screened Rutherford distribution is fitted to the forward scattering half‐angle of the Mott distribution and the size of the isotopic distribution is then adjusted to give the correct forward to backscattering ratio. This differential form has been used in a Monte Carlo simulation of the backscattering from Au. Both the total and the differential cross sections are straightforward to implement in a Monte Carlo simulation of electron/target scattering.

ISSN:0734-211X    

DOI:10.1116/1.585848

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

Techniques for doing alignment and registration in projection electron lithography are reviewed. The issues associated with extending these techniques to the case of high incident electron energy exposure are discussed. Measurements of backscatter electron emission contrast from W markers on Si wafers at 200 kV show that mark detection within the expected range of voltages is feasible.

ISSN:0734-211X    

DOI:10.1116/1.585849

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

For accurate linewidth measurement in scanning probe metrology the shape and size of the probe tip must be known. Since the probe can be degraded during a scan, quickinsitucharacterization is desirable. A technique is described employing an array of known structures that allows tip characterization with the probe microscope itself. This technique can be used to measure either the shape of a probe tip or the flexing caused by attractive forces near a sidewall. The results suggest that the sharpest probes may experience significant bending in the vicinity of a wall.

ISSN:0734-211X    

DOI:10.1116/1.585850

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

A low voltage backscattered electron collection technique has been developed. This technique collects low energy backscattered electrons in a sequential retarding field environment. This collection approach provides most of the desirable features for insulating package substrates and passivated integrated circuit (IC) devices inspection. These features are (1) high current and resolution, (2) high detector gain at low voltage, (3) good signal immunity from sample surface charging, and (4) on‐line topographical and compositional information at low voltage. In this paper, various details of the technique; such as detector arrangement, signal collection efficiency, sample charging immunity, beam current and resolution, image distortion and its comparison with conventional low voltage scanning electron microscopy inspection for insulating package substrate, and passivated IC inspection will be discussed.

ISSN:0734-211X    

DOI:10.1116/1.585851

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

We measure the fluorescent alignment generated by ion bombardment of an SiO2wafer mark scanned behind a corresponding window pattern in a silicon stencil mask. We conclude that an optimized system can align to 50 nm (mean+3σ) in less than 300 ms. Throughput is shown to be limited not by the alignment system, but by thermal loading of the mask during exposure.

ISSN:0734-211X    

DOI:10.1116/1.585852

出版商:American Vacuum Society    

年代:1991

数据来源: AIP

摘要:

In wafer alignment, subpixel resolution methods have been used in an attempt to increase the accuracy of the wafer positioning algorithm. These methods produce line positions that are between the pixels of the rasterized image of the alignment pattern. In this paper we provide a lower bound on the error variance of any alignment algorithm for a given alignment system and develop a method of ranking alignment patterns for their suitability for the task. We show that the use of a nonsymmetric alignment pattern will cause a decrease in accuracy unless information about the amplitude of the alignment pattern is known. An expression for the maximum useful subpixel resolution of any alignment system is derived and it is shown that with pixel resolution higher than this maximum value the subpixel methods produce no significant increase in accuracy. Examples of the use of the bound will be given.

ISSN:0734-211X    

DOI:10.1116/1.585853

出版商:American Vacuum Society    

年代:1991

数据来源: AIP