Electron Beam Focussing

The invention relates to focussing arrangements for focussing electron beams, such as are used in scanning electron microscopes or electron beam microfabrication apparatus for example.

According to the invention, there is provided a focussing method for measuring the focus of an electron beam, comprising the steps of scanning the beam across a predetermined and sharp-edged target for a plurality of times each at a different level of focus, producing an output signal during each such scan in dependence on the effect on the electrons in the beam of the said target, integrating the square of each such signal, and comparing the values of the integrated outputs to determine the value thereof corresponding to optimum focus.

According to the invention, there is also provided a focussing arrangement for measuring the focus of an electron beam, comprising scanning means for scanning the electron beam across a predetermined and sharp-edged target for a plurality of times, focus adjusting means for adjusting the focus of the beam to a different level during each such scan, video signal generating means responsive to the electrons in the beam as they are affected by the target during each such scan to produce a respective output signal during that scan, circuit means for squaring and integrating each such output signal to produce a respective integrated output, and comparing means for comparing the values of the integrated outputs to determine the value corresponding to optimum focus.

According to the invention, there is further provided a focussing arrangement for measuring and adjusting the focus of an electron beam and incorporated in apparatus having an electron beam source, scanning means for scanning the electron beam across a target area, focussing means for adjusting the focus of the electron beam, and signal generating means responsive to the electrons received in the target area to produce a corresponding electrical video signal, and comprising means operative to increment the focus level of the beam through a predetermined series of values thereof, a predetermined sharp-edged target in the said target area, control means operative to cause the scanning means to frame-scan the electron beam over the said target so as to perform a respective frame scan for each setting of the focus level of the beam, circuit means operative during each frame scan to derive, for each line scan thereof, an integrated signal corresponding to the integral of the square of the video signal produced by that line scan, circuit means operative to determine the maximum integrated signal for each frame scan, storage means operative to store each said maximum integrated signal and the corresponding value of the focus level, and comparing means operative to compare the stored maximum integrated signals whereby to determine the greatest one thereof and thereby the optimum value of the focus level.

Arrangements embodying the invention and for measuring and adjusting the focus of an electron beam will now be described, by way of example only, with reference to the accompanying diagrammatic drawing in which:

FIG. 1 is a diagrammatic view of a target which is scanned by the electron beam during the focus measurement;

FIG. 2 is a waveform of a video signal produced during the scanning process;

FIG. 3 is a block diagram of the focussing arrangement; and

FIG. 4 is a more detailed diagram of part of the block diagram of FIG. 3.

The focussing arrangement now to be described may be incorporated into a scanning electron microscope or electron beam microfabrication apparatus, in each of which, in use, an electron beam is directed at a target for inspecting or processing it. Clearly, for optimum operation, the electron beam needs to be correctly focussed, that is, the electron beam needs to be of minimum cross-sectional size at the working area.

The focussing arrangement carries out automatic adjustment of the focus current of the apparatus, that is, the current through the focus coil which determines the focussing of the beam, and monitors the crosssectional area of the beam at the working area, that is, the position where the target is located in use, so as to determine the focus current giving the best focus. The focussing arrangement may then automatically set the focus current to the optimum value.

The focussing arrangement carries out this process by scanning the electron beam across a target of predetermined cross-sectional area and having sharply defined edges. FIG. 1 shows the target 5 which in this example is circular in cross-section. During the focussing process, as will be described in more detail below, the electron beam is scanned a number of times across the target as shown diagrammatically by the lines 6 in FIG. 1.

FIG. 2 shows the video signal produced by one such scan, that is, the signal produced by the electron beam as affected by the target. As shown in FIG. 2, the level of the video signal drops abruptly (as shown in the region A) when the scanned beam is interrupted by the edge of the target and falls to a low level which is held until the beam reaches the opposite edge of the target and rises rapidly again, as shown in the region B in FIG. 2. The dotted lines in the regions A and B of FIG. 2 show the shape of the signal which is produced in conditions of optimum focus, that is, when the cross-sectional area of the electron beam is at a minimum where it strikes the target. The full line in each of the regions A and B show the video signal produced in conditions of poor focus; here, the beam cross-section is not at a minimum and therefore the beam is not abruptly cut off when it reaches the edge of the target during the scanning process but undergoes a comparatively gradual reduction until the whole of the beam area is obscured by the target. A corresponding process clearly occurs when the beam reaches the opposite edge of the target. In a manner to be explained, the focus arrangement sets the focus current at a particular level, close to one end of its range, and carries out a raster scan of the target (as shown in FIG. 1), monitoring and processing the video signals produced by each line scan of the raster. The focus current is then incremented to a new value and the sequence repeated, and so on until the focus current has been incremented to the opposite end of its range. The signals are then processed to derive the optimum focus current.

More specifically, it will be apparent that the video signal, V(x, y), derived from the target is a function of the target contrast g (x, y), the radius R of the beam spot at the target, and the subsequent signal processing, where x and y are the co-ordinates of the target surface. If the beam spot has an intensity distribution h (x.sub.o, y.sub.o) (that is, the intensity of the beam spot at a point x.sub.o, y.sub.o from its origin is h (x.sub.o, y.sub.o)), the signal produced when the beam spot is positioned at x, y, is the sum of the signals from all points which the beam spot illuminates and is given by

The two functions g(x, y) and h(x, y) can be represented by their 2-dimensional Fourier equivalents G(.omega..sub.x, .omega..sub.y) and H(.omega..sub.x, .omega..sub.y). Therefore Equation (1) can be rewritten as

This shows that the specimen contrast G is `filtered` or `modulated` by H. For perfect fidelity of imaging, .omega..sub.x, .omega..sub.y)=1.

The intensity distribution can conveniently be represented as ##EQU1##

Equation (3) can be rewritten in the form of its Fourier equivalent

Equation (4) shows that as R becomes smaller (that is, as the focus is improved), H increases for all nonzero values of .omega.. Substituting Equation (4) in Equation (2),

Therefore, V(.omega..sub.x .multidot..omega..sub.y) increases at all non-zero frequencies as the focus improves. Thus, by monitoring the video signal and measuring its total non-zero frequency content, a measure of the degree of focus is obtained.

In fact the specimen contrast G(.omega..sub.x, .omega..sub.y) has a large value at zero frequency. This in turn means that V(.omega..sub.x, .omega..sub.y) has a large value V(0, 0) at zero frequency. Since G(0, 0) does not depend on focus, neither does V(0, 0) and unless V(0, 0) is removed, the use of V(.omega..sub.x, .omega..sub.y) at non-zero frequencies as a measure of focus will be an insensitive measure.

Therefore, as will become apparent, the system to be described first measures the average value of the video signal for each line scan. This is equivalent to determining the zero frequency value of V(.omega..sub.x, .omega..sub.y) in Equation (5). This average value is then subtracted from the video signal and the result is squared and integrated, thus measuring the total non-zero frequency content of V(.omega..sub.x, .omega..sub.y). This is therefore a measure of the degree of focus b.ut without the insensitivity caused by the swamping effect of the contrast value at zero frequency.

In a manner to be explained, the various values of V(.omega..sub.x, .omega..sub.y) obtained in this way are compared to find the focus current giving the greatest value of V(.omega..sub.x, .omega..sub.y).

As shown in FIG. 3, the apparatus has a time base and scan generator 10, which may be of conventional form, and energises the normal "X" and "Y" deflection coils 12 and 14 to cause the electron beam to scan across the target. During this process, a digital-to-analogue converter 16 sets the focus current through the focus coil 18 to a predetermined level close to one end of its range.

The focus measuring arrangement 20, to be described in more detail, receives the video signal on a line 22 as produced by the electron beam (via a video signal generator 23) during the raster scan, and is controlled by line sync and frame sync signals received from the scan generator 10 on lines 24 and 26. The generator 23 may have any convenient form (e.g.a video camera), in dependence on the type of apparatus in which the electron beam is being used, so as to generate the video signal shown in FIG. 1. For each raster scan, the focus measuring arrangement 20 produces a focus measure signal on a line 28, dependent on the degree of focus, and this is converted into digital form by an analogue-to-digital converter 30 and stored in a memory 32.

Each frame sync signal is passed on a line 34 to increment a pre-settable counter 36 which energises the digital-to-analogue converter 16 to increment the focus current through the focus coil 18 to the next value, ready for the next raster scan. In fact, as will be explained in more detail below with reference to FIG. 4, each line is scanned twice in each raster.

A computer 38 receives the stored focus-representing digital signal in the memory 32 via a computer bus 40 when a complete series of raster scans has been carried out (that is, when the focus current has been incremented throughout its range). It processes these signals to determine which value of focus current gives the optimum focus. It then pre-sets the counter 36 accordingly so as to set the focus current at this value.

The focus measuring arrangement 20 of FIG. 3, and the manner of its operation, will now be explained in more detail with reference to FIG. 4.

With the focus current set to a particular value by the digital-to-analogue converter 16 (FIG. 3), the scan generator 10 causes the electron beam to perform the first line scan across the target 5 (FIG. 1) and the resultant video signal arrives on line 22 (FIG. 4) and is applied by a line 41 and a switch 42 to an integrator 44 having a resistor 46, an integrating capacitor 48 and a re-setting switch 50.

Switches 42 and 50 are controlled via a line 51 by control logic 52 which receives the line and frame. sync signals on lines 24 and 26 (see FIG. 3) from the scan generator 10.

During the first line scan, switch 50 is held open by the control logic 52 and switch 42 closed, and the integrator 44 therefore integrates the video signal. If the incoming video signal is represented by f(t), the output of integrator 44 at the end of the first line scan will be the mean value of f(t) for the line, that is, f(t).

At the end of the first line scan, the control logic 52 receives a line scan signal on line 24 and opens switch 42. Switch 50 is maintained open and therefore the integrated output of the integrator 44 is held. As previously explained, the scan generator 10 (FIG. 3) causes the line to be re-scanned (with the focus current unchanged).

During the second scan of the first line, the output f(t) from integrator 44 is passed through a resistor 54 to a point 56 where it is subtracted from the incoming video signal f(t) on line 22 which is applied to point 56 via a resistor 58. An amplifier 60 amplifies the difference signal so as to produce an output signal on a line 62 given by

The signal S.sub.1 is squared by a squaring circuit 64 so as to produce a signal

on a line 66 which is applied to a second integrator 68 having a resistor 70, an integrating capacitor 72 and a re-setting switch 74.

Switch 74 is controlled by the control logic 52 and is open during this second scan of the first line and therefore integrates the signal S.sub.2 to produce a signal ##EQU2## and this is passed to a peak detector 76 on a line 78.

At the end of the second scan of the first line, the line sync signal on line 24 causes the control logic 52 to close the switches 50 and 74. Switch 74 is maintained closed but switch 50 is immediately re-opened. Switch 42 is then closed and the first scan of the second line now commences (the focus current still being unchanged). Integrator 44 therefore now produces the mean value f(t) of the video signal for the second line. At the end of the first scan of the second line, the control logic 52 opens switches 42 and 74 (switch 50 being maintained opened).

The second line is now re-scanned (the focus current being unchanged).

Therefore, signal S.sub.1 is now produced on line 62 for the second line, squared by the squarer 64 to produce the signal S.sub.2 and integrated by integrator 68 to produce the signal S.sub.3.

At the end of the second scan of the second line, the control logic 52 once more closes switches 50 and 74, re-opens switch 50 and closes switch 42, and the sequence described above repeats (the focus current again being unchanged).

The peak detector 76 measures the peak value of the signal S.sub.3 at the end of the second scan of each line, and at the end of the frame is therefore holding the peak value of the signal S.sub.3 for the whole frame. When the control logic 52 receives the frame sync signal on line 26, it sends a control signal to the peak detector 76 via a line 80 which causes the peak detector 76 to output its stored peak value to the analogue-to-digital converter 30 (see also FIG. 3) on line 28. The peak detector 76 is then re-set and the signal on line 80 causes the analogue-to-digital converter 30 to convert the peak value into digital form and to pass it to the memory 32.

As already explained in connection with FIG. 3, the digital-to-analogue converter 16 then increments the focus current and the process described above with reference to FIG. 4 is repeated for the second frame scan.

During the first line scan of each line, switch 74 is held closed by the control logic 52 as already explained, and this therefore prevents the integrator 68 from producing an output.

When a sufficient number of frame scans has been carried out to have incremented the focus current throughout its range, the memory 32 will be storing a series of values of the signal S.sub.3, each value being the maximum value of that signal for a series of line scan carried out at a particular setting of the focus current.

The computer 38 then reads the stored values of the signal S.sub.3 and processes them to locate the value of focus current giving the maximum value of S.sub.3. For example, the computer may carry out a least squares curve fitting analysis on the stored values of signal S.sub.3 versus focus current.

The computer then sets the focus current to this optimum value via the counter 36 and digital-to-analogue converter 16 (FIG. 3).

Various modifications are possible in order to implement the process described. It may not be essential to carry out a frame scan at each value of focus current: a single line scan at each value of focus current may suffice. However, carrying out a frame scan at each value of focus current and detecting the peak value of the signal S.sub.3 during that frame scan overcomes errors which may be produced as a result of imperfections in the alignment of the electron beam column which can cause a shift in the scan position when the focus coil drive current is changed. Such a shift can cause a change in effective cross-section and this causes a change in value of the signal which is not related to an actual change in focus.