Recognition of diffraction-grating profile using a neural network classifier in optical scatterometry

Abstract

Optical scatterometry has been given much credit during the past few years in the semiconductor industry. The geometry of an optical diffracted structure is deduced from the scattered intensity by solving an inverse problem. This step always requires a previously defined geometrical model. We develop an artificial neural network classifier whose purpose is to identify the structural geometry of a diffraction grating from its measured ellipsometric signature. This will take place before the characterization stage. Two types of geometry will be treated: sinusoidal and symmetric trapezoidal. Experimental results are performed on two manufactured samples: a sinusoidal photoresist grating deposited on a glass substrate and a trapezoidal grating etched on a SiO2 substrate with periods of 2 microm and 0.565 microm, respectively.

Dimensional characterization of biperiodic imprinted structures using optical scatterometry

“…This step, generally called the inverse problem, can be solved using different approaches [11,13,[18][19][20][21]. The geometrical parameters are fitted using the Levenberg-Marquardt regression algorithm.…”

“…This latter has been successfully employed for 1D gratings and imprinted structures [10][11][12][13][14], and rarely for 2D patterns [15,16]. To the best of our knowledge, optical scatterometry has not been applied yet for biperiodic imprinted structures characterization.…”

Time-and-frequency domains approach to data processing in multiwavelength optical scatterometry of dielectric gratings

“…The principal task of scatterometry is detection and control of the deviation of a grating profile shape from the required one in grating manufacturing. Conventionally, it is performed by comparing the measured diffraction patterns (scattered field or other characteristics) with the theoretically calculated ones for a grating with desired profile, and subsequently determining whether the deviation is acceptable or not [1][2][3][4][5]. The technical facilities of modern lithography require theoretical models for a great variety of gratings profiles and materials.…”

“…The profound studies in practical scatterometry [1][2][3][4][5]22,23] assert that the simulation of measurement processes is very important not only for creation of the lookup tables facilitating the visualization process. The mathematical modeling oriented to extensive study of certain structures can result in the theoretically proven choice of the most efficient schemes for measuring.…”

Performance analysis of coherent optical scatterometry

“…This means that the problem is not only to regularize the inversion from the far field to the object, but also that the far field is only partly known. On the other hand, the importance of the technique in current industry processes is so high to deserve ad hoc studies in order to evaluate its real potentialities and limitations [4][5][6][7][8]. As an example, optical scatterometry is used in some applications of critical dimension (CD) metrology to determine with great precision (ranging from a few nanometers to even a fraction of a nanometer) the profile of a diffraction grating.…”