DE3228813A1 - Method and arrangement for determining structural edges in electron beam apparatuses - Google Patents

Abstract

A method with an associated arrangement for monitoring microelectronic structures with respect to position, width or distance as well as quality is applied in electron beam apparatuses in the field of microlithography. In this case, the measurement and monitoring of such structures of different size and shape down to structural widths and distances in the submicrometer range are to be enabled in a highly productive as well as automatic and objective fashion, the aim being to use, for the evaluation, a number of electrons that is sufficiently high for the accuracy requirements. According to the invention, a line is firstly multiply scanned perpendicularly to the edge direction, and secondly a plurality of lines are singly scanned at respective prescribed distances along the edge direction. A signal curve which is indicative of the shape of the edge is formed in each case for the first and second cases by summation of individual signal curves. The edge roughness is determined from the gradient of the two curves. In order to carry out the method, the electrons emerging from the object upon impingement of the electron beam are fed to a receiver of the circuit arrangement according to the invention.

Description

Method and arrangement for determining structure edges in

Electron beam devices The invention relates to a method and an arrangement for its implementation, the control with regard to position, width or the distance and quality of microlithographically produced structures are used, by automatically determining the course of the structure edges.

The invention can be used in the field of microlithography ie in Electron beam devices used to measure and control such structural patterns to serve. The structures can be used as adhesive paint masks, in metalized glass bodies or in the form of semiconductor wafers.

It is known to use light-optical measuring devices for such measurements, as from prospectuses of the Pa. Nikon / Japan to "Micro-Pattern Inspection Station" or to "Laser Interferometric x-y Measuring Machine" states.

However, these methods are only possible as long as the minimum element dimensions or distances are not less than a few micrometers.

For reasons of intensity, a light beam with a rectangular Cross-section, its narrow side up to 1 µm and its long side up to a few 100 pm tears, led over the edge of the structure and the transmitted or reflected Light signal evaluated by photo detectors.

The disadvantage is that this makes demands on the geometric shape of the structures to be measured are connected and thus the variety is significantly limited will. It is also known to use electron beam devices for the control of structures to use. For example in the patent specification of DD -,; P 124 091 is described that on the principle of a scanning electron microscope a fine electron probe guided over the object to be measured, an image of the object synchronously with it displayed on a screen and compared with a projected scale network will. The measurement is carried out directly with the aid of the scale network or through defined Object displacement using a reference mark projected on the fluorescent screen.

The disadvantage is that the measurement productivity is too low and the Measurement accuracy is subjectively influenced.

The aim of the invention is to provide a way of measuring and controlling of microlithographically produced structures of different sizes and shapes to create structure widths and spacings in the sub-micrometer range. The procedure should be highly productive and run automatically and objectively.

The invention is based on the object of such a method to develop with an arrangement for its implementation that allows, on the one hand to carry out the measurement of structure widths in the submicrometer range, on the other hand a sufficient number of electrons for the accuracy requirements Based on the evaluation, as well as obtaining statements about the quality of the structure edges.

In a process for determining the course of structure edges in measuring devices to determine the position, the extent or the distance as well as the Quality of preferably microlithographically produced structural patterns, the signals for detection resulting from an atracing with the electron beam the structure edges are exploited and the electron beam gradually over the edge to be measured is guided, the object according to the invention is based on solved the following process steps: a) In the vertical to the hantenaas expansion The sweep direction becomes the structure pattern along a line in constant predetermined steps by shifting the electron beam scanned point by point, which in each Point received signal is digitized and stored in the memory of a computer. If the sea through the electron probe is small, then only the property becomes the edge detected in this small area and on the other hand also no requirement raised to the length of the edge.

b) Step a) is repeated several times (n times), with the coordinate of the sweep in the direction of the edge is not changed in order to reduce the signal-to-noise ratio to improve the signal obtained in a).

c) The digitized signal values obtained in steps a) and b), which belong to the same locations along the sweep direction become a signal curve summed up.

d) For n different locations at given distances a along the structure edge, which is perpendicular to the Sweepriolitung, step a) is repeated in each case, to reduce the influence of the edge roughness on the measurement signal.

e) The scanning signals obtained in step d) are digitized and summed for all sampling points of the same coordinate in the sweep direction. That so The signal obtained turns out to be a spatial-temporal mean. an edge area of n.a. The distance a is chosen so that the sweep offset of the edge extension is adapted, i.e. n.a # 1, 1 = edge extent.

f) When the signal is generated in step e), the edge roughness is finite Signals superimposed on one another in the sweep direction in the order of magnitude of the edge roughness are shifted. The resulting edge signal is therefore generally flatter as one in which a = - 0 holds.

The steepness of the signal curve obtained according to c), i.e. for a = O, is divided by the steepness of the signal curve obtained according to e), i.e. for a # O and this quotient is linked to a constant determined from test objects. This results in a value that is a measure of the roughness of the edges to be measured Structure is.

An arrangement is used to carry out the method according to the invention in an electron beam device, in which the metal electron beam is aimed at the Structure impinges and depending on this structure an exiting Electrons formed signal is fed to a receiver.

The arrangement is characterized by the following features: The receiver signal, which can be very different in size and a disturbing base level contains is fed to a normalization element. The normalized signal is in a AD converter that is read in from the central processing unit of a computer via a gatekeeper is digitized and fed to the central unit.

For digital displacement of the electron beam for the x and y coordinate directions a position counter is provided for each, which uses the D / A converter and power amplifier Position of the electron beam determined by a deflection system. At an entrance Each position counter contains information about the start and end values in the sweep direction at. When the end value of a position counter is reached, a message is sent to the central unit. At the same time the position counter is reset to the initial value and a clock pulse is outputted to the position counter of the second coordinate direction. As a result, a new scanning sweep can take place on a line that is parallel to the preceding runs, with the two lines having a predetermined distance a.

There is an electronic one between the two position counters Switching device for selecting the scanning sweep in the x or y coordinate direction. Included The read-in pulses of the signal digitization serve as clock pulses, which also can be fed to the AD converter as a start signal via a delay element.

The read-in pulse transfers the digitized signal to the central unit accepted. The trailing edge of this pulse increases the position counter and the electron beam takes the next position. After a delay, the adapter will started and digitize the new signal value To smooth out noisy Signals is a low-pass filter connected in front of the normalization element, its time constant is matched to the delay time of the AD converter.

To improve the edge signal in terms of evaluability can it can be advantageous to add or subtract the signals from different receivers to mix.

To improve the accuracy of electron beam positioning The following features of the circuit arrangement are used: The adjustment of the step size to the structure to be measured and the deflection sensitivity of the deflection system via a pulse multiplier stage and a digitally adjustable analog attenuator.

The inclination of the object against the direction of the blenksysteme is taken into account by rotation correction units.

The position counter with step size adjustment and fast DA converter is only symmetrical to the assumed canter position for the actual raster sweep used, and the basic deflection of the beam to this assumed edge position set by a second DA converter before the start of the series of measurements. This DA converter is more accurate and slower, its output is the rasterized deflection voltage additively added.

In the following, the invention will be based on an exemplary embodiment and the accompanying drawing.

In the drawing, FIG. 1 shows the course of the structure edge, FIG. 2 circuit arrangement for detecting structure edges, FIG. 3 circuit arrangement to improve the edge signal, Figure 4 circuit arrangement for improvement electron beam positioning.

The solution according to the invention for detecting structure edges works assume that an electron beam is digitally shifted and that from a receiver received signal is digitized.

FIG. A shows a structure 1 on a carrier material in cross section 2.

In Fig. Ib the structure edge 3, which is to be measured, is shown. The y-coordinate direction was chosen for the position of the structure edges. Perpendicular in addition, the sweep direction is in which the structure edge 3 in predetermined steps, e.g. 5 nm, is scanned. The signal obtained from backscattered electrons at each sampling point, Secondary electrons or a mixture of both is digitized and stored. Since the diameter of the electron probe is small, i.e. typically 10 nm only the edge in this small area is recorded, on the other hand also none Requirement for the longitudinal expansion of the edge. The dwell time per point is e.g. 4 µs. Due to the small number of electrons, it registers the signal 1c shows a small signal-to-noise ratio.

In order to improve this, the sampling at y = constant becomes n times (e.g. n = 50) repeated and the digitized signal values to a signal curve 5, which is shown in Fig. 1d. In this example the Detection of the edge n.2ms = 0.1s. With this type of signal formation, the edge roughness is not taken into account.

Therefore, after the btastsweep at y1, after each shift of the electron beam along the y-direction by dy n further measurements are carried out. The resulting signals are superimposed to form a signal curve 6, which is shown in FIG. 1 f is shown and a spatial-temporal mean over the edge area n. #y represents, whereby dy is chosen so thatn. #y # 1 (1 = edge expansion) applies. the The steepness of the signal curve 5 is divided by the steepness of the signal curve 6 and this quotient with a constant linked that from test objects was determined. This creates a value that is a measure of the edge roughness.

Figure 2 shows a circuit arrangement for performing the method, in which a signal from a receiver 7 in a normalization element 8 by additive Level shift and multiplicative gain change is normalized. To digitization this signal is used by an AD converter 9, which via a gate circuit 10 from the central unit 11 of a computer is read. A delay element 21 receives the read-out pulse EI from the central unit 11 and emits a start signal St to the AD converter 9. The digital shift is determined by a position counter for each coordinate direction 12 and 13, each via a DA converter 14 and 15 as well as a power amplifier 16 and 17 a deflection system 18 for the di: e x direction and a deflection system 19 for the influence y-direction. The start and end values Aw and 3w of a raster sweep become from the central unit 11 to the position counters 12; 13 given, they describe the presumed edge position. By an electronic switching device 20 can optionally an x or y line can be generated. The clock pulses are used Read-in pulses EI of the signal digitization. When reaching the end value of a line, which lies in the x-direction in the example, a signal is generated by the position counter 12 M given to the central unit 11. At the same time this position counter is on the initial value is reset and a clock pulse is sent to the position counter 13. A renewed hbtastoweep takes place on a line that is parallel to the first and the a stand #y is offset.

FIG. 3 shows a circuit arrangement for improving the edge signal shown in which two receivers 22 and 23 for backscattered electrons and a receiver 24 are provided for secondary electrons. The output signals of the receivers 22 and 23 are mixed additively or subtractively in a summing element 25, the resulting Signal and the output signal of the receiver 24 for smoothing each over a Low-pass filters 26 and 27 are fed to the normalization elements 28 and 29, their output signals linked additively or subtractively in a further summing unit 30 and in one AD converter 31 are digitized.

The time constants of the low-pass filter 26 and the low-pass filter 27 are on the delay time of the HD converter 31 matched.

The circuit arrangement of FIG. 3 brings about an improvement Electron beam positioning for carrying out the method according to the invention. The adaptation of the distance dy to the structure edge to be measured of length 1 as well as the deflection sensitivity of the deflection systems 18; 19 takes place via a pulse multiplier stage 33, which is connected upstream of the position counter 13. There is also an analog for it Attenuator 35 is provided, which is connected downstream of the DA converter 15.

analogously to this, the step size in the x-direction is determined by a pulse multiplier stage 32, which is connected upstream of the position counter 12, and by an analog attenuator 34, downstream of the Driandler 14, realized.

The inclination of the structure to be measured against the direction of the deflection systems 18; 19 is through rotation correction units 36; 37 taken into account. An adjustable Part of the deflection voltage of one coordinate additive or subtractive to the other Coordinate added.

win a considerable increase in accuracy in the positioning of the Electron beam is achieved in that the position counter 12 with the fast DA converter 14 and the attenuator 34 only for the actual sweep, which in the example takes place in x-Xichtuna and symmetrically to the edge position. The basic deflection of the electron beam to this assumed edge position is through a second DA converter 38 or 39 for the x or y coordinate direction before the start of the series of measurements set.

These work more precisely and more slowly, e.g. with 16 bit.

The digital value of this basic deflection is provided by the central unit 11 determined and in two of the DA converters 38; 39 upstream storage tanks 40; 41 issued. The output signals of the DA converter 38; 39 are each with the output signal of the D / A converter 14 or 15 of the corresponding coordinate in a summing element 40 or 41, which also includes the output signal of the relevant Rotation correction unit 36 or 37 is supplied and that with one of the power amplifiers 16 or 17 is connected.

To achieve a high level of automation, it is beneficial to all adjustable units digital, i.e. marked with d in Figures 3 and 4, to be controlled by the central unit 11 and additional feedback to the central unit to form, e.g. maximum and minimum value of the standardized signal, position of the position counter and readiness of the AD converter.

By setting the pulse multiplier 33 accordingly, the Spacing # y between lines can be varied or the same line repeated several times Maintain scanning.

If a further enlargement of au is necessary, the clock pulses are fed into a higher-order bit position of the position counter. The LT announcement to the central unit 11 about the presumed cold position can be done by the Sollpoaition is entered into the central unit 11. This procedure is used to Measurement of known structures at given measuring points. On the other hand, the ..notification takes place by presenting the object on a screen and the edge to be measured is marked in a suitable manner. For image-generating screening the position counter 12; 13 used. The marking can be e.g. a faded in, movable mark.

L e r s e i t e

Claims (5)

Claim: 1. Method for determining the course of structure edges in measuring devices to determine the position, the extent or the distance as well as the Quality of structural patterns preferably produced by microlithography, whereby the Signals generated when scanning with the electron beam to detect the structure edge be exploited and the electron beam gradually over the to be measured ante is driven, characterized by the following procedural steps: a) in the vertical The Swesprichtung standing for the edge expansion is along a line in constant predetermined steps the structure pattern is scanned point by point by means of an electron beam, the scanning aials are digitized and stored, b) step a) is repeated repeated, the coordinate in the direction of the edge not being changed, c3 die in steps a) and b) obtained and digitized signal values, respectively belong to the same locations along the sweep direction are added to a signal curve, d) for n different locations at given distances a along the structure edge, perpendicular to the sweep direction, step a) is repeated in each case e3 in step d) obtained and digitized scanning signals become the same for all scanning points Coordinates are summed up in the sweep direction, and signals are given as spatio-temporal Average values formed over a selectable structure edge area n.a., f) the steepness of the signal curve obtained according to c) is determined by the steepness of the e) obtained signal curve divided, and this quotient with one from test objects determined constants linked to a value that is a measure of the edge roughness of the structure to be measured. 2. Arrangement for performing the method according to claim 1, in one Electron beam measuring device in which the metal electron beam hits the structure to be measured and, depending on this structure, an electron emitted from it formed signal is supplied to a receiver, characterized in that the Receiving signal via a normalization element and an AD converter, which is via a gate circuit is read in by the central unit of a computer, fed to the central unit is that a read-in pulse is sent to the AD converter via a delay element as a start signal is supplied that for digital displacement of the electron beam for the x; and y-coordinate direction a position counter is provided that at an input Each position counter has preset start and end values of the raster sweep, that in each case a first output signal of the position counter via a DA converter each and a power amplifier, a deflection system for positioning the electron beam is fed in the relevant Koordins tenrichtung and each a second output each position counter is connected to the central unit, these two output signals Information about reaching the end value in the sweep direction contains that between Both position counters have an electronic switching device to select the sweep direction is provided in the x or y coordinate direction, the read-in pulses as clock pulses serve to digitize signals, that the time outputs of the position counter are each connected to the other position counter via the switching device. 3. Arrangement according to claim 2, characterized in that several Receivers are provided, each of which has a delay element acting as a low-pass filter are twisted with a normalization element, and the output signals of all normalization elements supplied via an additive or subtractive summing element with the AD converter are. 4. Arrangement according to claim 3, characterized in that at least a part. the receiver is followed by a summing element and this with a Delay element is connected. 5. Arrangement according to claim 2, characterized in that for adaptation the step width in the sweep direction and the distance a in the direction of the edge to the to measuring structure in the clock input line of the position counter an adjustable Pulse multiplier and each between the DA converter and the power amplifier an adjustable analog attenuator is connected that to correct the Inclination of the object Rotary correction units are provided that each coordinate direction a second, more precise and slower than the first working DA-; whose output signal is additively linked with the rasterized deflection voltage is.