SU838326A1 - Method of checking microobject linear dimensions - Google Patents

Description

(54) METHOD OF CONTROL OF LINEAR DIMENSIONS OF MICRO-OBJECTS

the size of the step of the microstructure and the width of its elements; in fig. 2 is a functional diagram of a device implementing the proposed method; in fig. 3 - the form of signals arriving at the computing device.

A device that implements the proposed method contains an optically coupled projection microscope 1, a movable mirror prism 2, mirror reflectors 3 and 4, a lens 5, a bifurcation circuit 6, a slit diaphragm 7 displaced by some distance from the central axis of the optical system, a photocell 8, the amplifier 9, the device 10 for moving the structure, the interference measuring transducer 11, the computing device 12, the recording device 13.

The method consists in creating a luminous flux, on the way of which a controlled periodic microstructure is placed and moving it in the direction of the measurement line, forming an enlarged optical image of the elements of the microstructure, bringing them closer to a distance equal to the nominal image size of the element (Fig. 1a, where the distance h is equal to d (the nominal size of the width of the element) and split the resulting image into two half-contrast ones with a shift relative to each other in the direction of the measurement line by an amount equal to the nominal size of the item. Due to this, in the ideal case, when there is neither an error in the step, nor an error in the width of the element of the microstructure, the optical image is an integral monochromatic semi-contrast (Fig. 1b). In the real case, when there are pitch errors and widens in the widths of the microstructure elements, either dark (Fig. 1g, g) or light (Fig. 1c, e) stripes are observed at the places of superposition or separation of semi-contrast images. The amount of overlap or separation of the semi-contrast images of the same element corresponds to the deviation of the actual size of the element width from its nominal value (Fig. D, g), the amount of overlap or separation of the semi-contrast images from the neighboring elements corresponds to the complex error, i.e. the sum of errors schaga structure and errors of the width of its elements (Fig. 1 c, d, e, f, g). During the movement of the microstructure, all the borders of the contact between the semi-contrast images by the photocell are sequentially recorded, the duration of the electrical output signal of which is generally proportional to either the element size errors or the complex error of the microstructure step. Allocation of the schaga fault from the complex error is carried out by algebraic subtraction from it of the parallelities of the widths of the microstructure elements recorded in parallel.

The method is carried out as follows.

Periodically, the microstructure 14 is placed in front of the lens of the microscope 1 on a moving carriage (not shown in Fig. 2), the movement of which sets the device 10, and illuminated with a light flux from a coherent source (not shown in Fig. 2) optical radiation. Using a microscope 1, they form an optical image of the elements of the microstructure, direct it to the movable prism 2, give it an initial displacement relative to the axis passing through the right angles of the mirror reflectors 3 and 4 so that the images of the elements are shifted to a distance equal to the nominal the size of their widths (Fig. 2 a). The prism 2 is shifted in the direction indicated by the arrow. For each type of microstructure, this displacement is strictly defined and does not change during the control process. The optical image of the structure elements formed in this way is then split in two by means of a split-up circuit 6 into two semi-contrast and displaced relative to each other in the direction of the measurement line by an amount equal to the nominal size of the width of the element. As the microstructure moves, there are changes in the illumination of the slit diaphragm 7, which are recorded by the photocell 8. The photoelectric pulses are amplified by the amplifier 9 and fed to the computing device 12. The frequency of the pulses from the interference measuring transducer 11 is proportional to the speed of movement, and the "price of the pulse corresponds to one period interference stripes. These pulses are used to quantize the period of the pulses coming from the amplifier 9. The device 13 records the calculated values of the error value of the microstructure step and the error of its elements in fractions of the quantizing pulses.

The duration of incoming rectangular pulses is proportional to the error of chiriya (Fig. 3) of the previous element Ad, the complex error Ak, and the error of the width of the subsequent element Adj ...

. Ak & h + A.4i +

whereDp - step error.

The selection of the error of a step from complex error may, for example, occur as follows.

The signs for all three faults (and for each subsequent triple) are determined.

The sign “- for D d corresponds to the separation of the semi-contrasts of the same element, and the sign“ + - respectively, their imposition. For Ak, on the contrary, the sign “-f corresponds to the separation of semi-contrasts

neighboring elements, and the sign “- - overlay. In the case where Ad | and Adz are positive, and Dk is negative, the polarity of the pulses indicates the sign: if Ad, and Adj are of the same sign, then the magnitudes of the errors are added, and if they are, they are subtracted. The positive error And d reduces And k, and negative - increases.

Thus, by simultaneously measuring and taking into account the errors in the widths of the microstructure elements and the distances between them without rebuilding the optical channels, the measurement accuracy is improved by controlling the steps and widths of the microstructure elements from one installation of the product.

AND

Claims (1)

Invention Formula The method of controlling the linear dimensions of microobjects on the author. St. No. 612148, characterized in that, in order to increase the accuracy of monitoring the microscopic microstructures, the images of the microstructures are biased prior to their splitting one into the other in the image plane by a distance equal to the nominal width of the microstructure element. Sources of information taken into account during the examination 1. USSR Author's Certificate No. 612148, cl. G 01 B 11/08, 1978 (prototype). 7 8