Radiation imagery chemistry: process – composition – or product th – Including control feature responsive to a test or measurement
Reexamination Certificate
2000-11-22
2002-11-26
Young, Christopher G. (Department: 1756)
Radiation imagery chemistry: process, composition, or product th
Including control feature responsive to a test or measurement
C430S944000, C250S472100, C250S473100, C250S474100
Reexamination Certificate
active
06485872
ABSTRACT:
BACKGROUND OF THE INVENTION
Many current processes for manufacturing microelectronic devices, and the like, entail the fabrication of stacks of thin films having compositions that must be controlled to satisfy exacting standards and extremely tight tolerances. For example, fluorinated silica glass (FSG) films are used as dielectric layers in semiconductor integrated circuits (IC's); because the fluorine constituent serves to reduce the dielectric constant of the layer, and thereby to improve circuit speed and reduce cross-talk between wires, it is important to control its concentration so as to optimize the electrical, chemical, and other properties of the layer. Similarly, the dopant concentrations in boron- and phosphorus-doped silica glass (BPSG) films, which are used as intermetal dielectrics, are quite critical and must be monitored and controlled carefully to maintain high yields in production, and the same is true of many of the newer materials that are now being employed as thin films, which materials tend to be compounds with complex chemical compositions and interfaces that must be monitored and controlled.
It is now also common for integrated circuits to include complex submicron geometric structures. For example, dynamic random access memory (DRAM) devices are typically formed with deep, tapered trenches etched into the silicon to thereby provide larger area capacitors without increasing the chip surface area utilized. The sidewall taper angle and depth of such trenches must be controlled carefully to maintain yield; moreover, the formed trenches may themselves be filled subsequently with materials such as dielectrics and polysilicon, giving rise to the need for yet further process monitoring and control capability.
In addition to the foregoing, compositionally graded structures are often employed in thin film manufacturing; dopants are typically diffused or implanted into semiconductor devices to provide diffuse composition profiles; and silicon on insulator (SOI) wafers are commonly fabricated by forming a buried oxide layer with a graded composition profile, which must be sharpened by thermal annealing. Controlling the depth profiles of such structures is of critical importance.
In all of the processes referred to, as well as in others that will be evident to those skilled in the art, it is necessary to measure the properties and characteristics of at least one layer of interest. Typically, and for any of a number of reasons, such measurements cannot presently be performed directly on the product wafers: i.e., product wafers with patterns, and multilayered film stacks, are often too complex for practical data analysis; measurements are often either destructive to the sample or are such that they pose a significant risk of contamination or defect introduction; and current techniques, carried out in the visible and ultraviolet wavelength ranges, are not sufficiently sensitive to enable accurate measurement of crucial layer properties, such as composition and carrier concentration. In any event, nondestructive measurement techniques will usually be preferred, largely because they enable the use of in-line process control methodologies without undue consumption of expensive, non-product test wafers.
SUMMARY OF THE INVENTION
Accordingly, broad objects of the present invention are to provide a novel method and apparatus for measuring and controlling the composition and other properties of thin films, in accordance with which at least certain of the limitations and deficiencies of current techniques and procedures, as described herein, are minimized, avoided and/or overcome.
It has now been found that certain of the foregoing and related objects of the invention are attained by the provision of a method for estimating at least one parameter of a sample, determined from the dielectric function of a material of which at least one layer of the sample consists. The method comprises the steps:
(a) providing a sample comprised of at least one layer and having a substantially specular surface;
(b) defining an optical model of the sample along a direction perpendicular to its surface and based upon reflectance values, the “at least one layer” being defmed in the model by a thickness value and, for each of a multiplicity of wavelengths in the infrared spectral region, by a dielectric function value;
(c) providing a training set consisting of measured values of the “at least one parameter” and an associated dielectric function, the measured values being obtained from a multiplicity of samples selected to represent a range of values of the at least one parameter;
(d) determining from the training set a predictive mathematical relationship between the at least one parameter and the associated dielectric function, so as to enable prediction of the at least one parameter from input values of dielectric function;
(e) irradiating the specular surface of the sample with infrared radiation, including the multiplicity of wavelengths referred to in step (b), and obtaining a measured reflectance spectrum composed of values obtained over the multiplicity of wavelengths;
(f) simulating a reflectance spectrum from the optical model at the multiplicity of wavelengths using various values of the dielectric function calculated from assumed dielectric function descriptors and a value of the thickness of the at least one layer, and computing the various values of the dielectric function descriptors so as to minimize the difference between the simulated reflectance spectrum and the measured reflectance spectrum, thereby determining an optimized dielectric function value for the at least one layer at the multiplicity of wavelengths; and
(g) calculating the value of the at least one parameter using the optimized dielectric function value and the predictive mathematical relationship.
In certain preferred embodiments a pattern of variation derived from the training set is utilized so as to constrain the number of the descriptors required to describe the dielectric function. The values of dielectric function used may be parametrized as weighted linear superpositions of vectors determined to span the space of dielectric functions derived from the training set, with the descriptors being the coefficients of the vectors. At least one of the vectors will desirably be determined through a multivariate statistical regression of the set of dielectric functions measured in creating the training set. The predictive mathematical relationship may for example be determined through a multivariate statistical regression of the training set; it may be determined employing a neural network algorithm calibrated with the training set; or it may be determined by establishing a library of dielectric function values with associated values of the at least one parameter, organized in the form of a look-up table which is accessed to determine the parameter from the optimized values, and access may include the additional step of interpolating between elements thereof. The predictive mathematical relationship may be established between the at least one parameter and spectral features derived from the dielectric function of the training set. The spectral feature may be at least one characteristic of at least one peak observed in the training set dielectric function, such a characteristic typically being the intensity, position, height or width of the at least one peak. In certain instances the thickness value will be varied in step (f) as well as using the computed values of the dielectric function descriptors.
The at least one parameter calculated by the present method may represent the species and/or the concentration of at least one chemical constituent of the material of the at least one layer. In particular, the parameter may be the concentration of fluorine atoms within the material, the concentration of hydroxyl groups within a dielectric matrix, the concentration of water molecules within a dielectric matrix, the concentration of hydrogen atoms within a dielectric matrix, or the concentrations of at least one of bor
Charpenay Sylvie
Rosenthal Peter A.
Yakovlev Victor A.
Dorman Ira S.
MKS Instruments Inc.
Young Christopher G.
LandOfFree
Method and apparatus for measuring the composition and other... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for measuring the composition and other..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for measuring the composition and other... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2923428