This invention relates to a charged particle beam system and more particularly to such a system wherein beam focusing is controlled by controlling the electric potentials distributed over the surface of the sample through the measurement of those electric potentials.
A scanning electron microscope (SEM) and a focused ion beam (FIB) apparatus are examples of an apparatus which uses a charged particle beam such as an electron beam or ion beam to inspect, measure and work a sample. In these apparatuses, the beam of accelerated charged particles is focused on the sample by controlling the retarding (decelerating) voltage applied to the objective lens or the sample; the charged particle beam is scanned in a pattern of raster; the secondary electrons emitted from the sample are detected in synchronism with the raster scan; and the two-dimensional signals (image) representative of the superficial shape of the sample can be obtained. The charged particle beam system of most recent appearance is provided with an auto-focus mechanism so as to obtain clear images.
If the sample is intensely charged and the amount of charge is unknown, the focal point shifts according to the amount of charge. Therefore, in effecting the auto-focusing function, it is necessary to expand the variable range of the exciting current for the magnetic objective lens or the retarding voltage. Consequently, the time required to complete the auto-focusing operation is lengthened, resulting in a problem that the measurement time is prolonged.
When the beam is to be focused on the surface of the sample, which is charged, through the control of the magnetic objective lens, the focal point shift due to the charge on the sample surface must be adjusted by adjusting the exciting current for the magnetic objective lens. In this case, however, if the magnification is calculated on the basis of the adjusted exciting current, the calculated magnification does not coincide with the true magnification.
Several methods have been proposed to solve the problem resulting from the electrification of the sample surface, that is, the problem that the auto-focusing operation is prolonged or that the calculated magnification is erroneous. Those methods are disclosed in, for example, patent documents WO2003/007330, JP-2001-52642, and WO99/46798. According to the document WO2003/007330, the amount of electrification of the sample is measured by the electrostatic capacitance measuring device when the sample is carried into the sample chamber. The document JP-2001-52642 discloses the procedure wherein the amount of charge on the sample surface is determined by using the electrostatic capacitance detector installed in the sample chamber. Accordingly, the sample is irradiated by the charged particle beam such as an electron beam having a desired kinitic energy while the retarding voltage is so controlled as to cancel those electric potentials over the sample surface which are measured by the electrostatic capacitance detector. The document WO99/46798 proposes a scanning electron microscope which can eliminate the adverse effects on the image due to the electrification of sample.
Indeed the above mentioned, electrostatic capacitance measuring device can rapidly measure the surface potential of sample, but it cannot measure the potential in a small area, i.e. the local potential, on the sample surface.
In order to measure the amount of the sample electrification with the system disclosed in the document JP-2001-52642, the sample must be carried near to the electrostatic capacitance detector. This result in a decrease in throughput and a need for an additional provision of a circuitry for controlling the electrostatic capacitance detector.
In the document WO99/46798, the electric potentials over the sample surface are controlled in such a manner that while the secondary electrons are being driven off the sample surface by the primary electrons irradiating the sample surface, the output of the secondary electron detector becomes maximum. In order to employ this method in the case where the sample surface is highly charged, the accelerating voltage for the primary electrons constituting the electron beam must be increased to cause the beam to reach the sample surface. This method, therefore, does not meet the need of today that measurement should preferably be made using as low an accelerating voltage as possible. Further, according to the document WO99/46798, since the condition in which the output of the secondary electron detector becomes maximum is searched for while the secondary electrons are being emitted, the damage to the sample surface due to the electrons hitting thereon is considerable and it is also difficult to determine the condition to give the maximum output.
Furthermore, it has come to be known that there exists the phenomenon of electrification which is different from the conventionally known “wafer electrification”. Namely, samples have come to be known which, in the most recent semiconductor manufacturing process, although there is no surface electrification of the wafer when the retarding voltage is absent, the application of the retarding voltage causes the wafer surface to be charged electrically. The cause of this type of electrification is still unknown, but the phenomenon is supposed to be attributed to the change in the surface potential due to the influence of electric or magnetic field existing at the very point of observation. In order to solve this problem, it is necessary to measure the surface potentials under the condition of electric or magnetic filed approximating the condition for actual observation.
The object of this invention is to provide a system for measuring the potentials over the sample surface rapidly with high precision while the retarding voltage is being applied to the sample.
The concrete methodology according to this invention is as follows.
When the retarding voltage is increased to decelerate the charged particles forming the charged particle beam and reaches the level at which the kinetic energy of the charged particles becomes nearly equal to 0 eV, the charged particles impinging on the sample surface can not give rise to secondary electrons. Use is made of this phenomenon, that is, the potentials over the sample surface are estimated by measuring the retarding voltage which reduces to 0 eV the kinetic energy of the charged particles reaching the sample surface.
Concretely described, the specific value of the retarding voltage which minimizes the output of the secondary electron detector is obtained between the value of the retarding voltage at which the charged particle beam does not reach the sample surface and the value of the retarding voltage at which the charged particle beam can reach the sample surface. Then the surface potential of the sample is determined depending on the specific value of the retarding voltage. Accordingly, the charged particle beam having a desired energy is cast onto the sample by controlling the retarding voltage so as to cause the determined surface potential to be offset.
According to the above described constitution, the potentials over the sample surface are measured, and the measured potentials are offset, whereby the charged particle beam having a desired energy can be cast onto the sample. Consequently, the magnification error does not occur, and measurement can be effected at correct magnifications.
Further, since the surface potentials are offset, the time required for auto-focus operation can be shortened. As a result, throughput can be improved for a charged particle beam system which is used to inspect patterns or to measure pattern dimensions.
Furthermore, since the charged particle beam is used, the amount of charge on and within a tiny area of the sample surface can be measured.
Other objects of this invention and other concrete structures according to the invention will be described in the following description of the embodiment of this invention.
A scanning electron microscope as a first embodiment of this invention will now be described with reference to the attached drawings.
Here, the scanning of electron beam by using the retarding voltage will be described briefly.
In the operation of a scanning electron microscope or a focused ion beam apparatus, the irradiation energy of the electron beam is controlled by applying a voltage to a semiconductor wafer so as to focus the beam of the charged particles on the wafer without damaging the elements in the wafer surface due to the beam irradiation. This voltage applied to the semiconductor wafer is called “retarding voltage”.
The scanning electron microscope produces the shape of the surface of a sample by scanning the electron beam on the sample surface by means of a scanning coil 6, transducing the output of the secondary electron detector 10 into a brightness signal in synchronism with the scanning, and displaying the brightness signal on the screen of a picture tube as the two-dimensional image of the sample surface.
According to this invention, the retarding voltage is controlled; the thus emitted secondary electrons are detected; the sample surface potential is calculated by a computer 11 according to the detected secondary electrons; the surface potential is estimated by a surface potential estimator 12; and the energy of the electron beam reaching the sample is adjusted to a desired value through the control of the retarding voltage.
To obtain a clear image of the sample surface, the beam of charged particles must be exactly focused on the surface. However, it is difficult to exactly focus the beam on the sample surface due to the roughness of the surface or the electric charge existing in the surface. Therefore, the charged particle beam system of today incorporates an auto-focus mechanism therein. According to many auto-focus mechanisms ever proposed, the focal point of the charged particle beam is changed by discretely changing the exciting current for the objective lens or the retarding voltage, and the images of the sample surface are taken for the several focal points. The images associated with the different focal points are processed with a focus evaluation filter (differential, second-order differential, Sobel, Laplacian, etc.) to produce the associated focus evaluation images leading to the calculation of the focus evaluation value (referred to also as sharpness). Here, the weighted sum, the average or the variance, of the focus evaluation images can be used as a measure of the focus evaluation value. These steps of operation are usually performed as an auto-focus operation. The value of that exciting current or retarding voltage which gives the maximum of the focus evaluation value is regarded as the exciting current or the retarding voltage that occurs when the beam is properly focused on the sample surface. Throughout this specification, the operation of changing the exciting current for the objective lens or the retarding voltage is called “sweeping” and the range within which they are swept is termed “sweeping width”.
It has been recently observed that some wafers remain kept at a certain fixed potential even after it is electrically grounded. This fixed potential is attributed to the polarization of polar material in the photoresist due to the friction generated during the process of resist coating by using the spin coater, or ascribed to the electrification due to plasma etching.
Especially in the case where the SOI (silicon on insulation) technique is employed, that is, where an insulation film is formed on a wafer and then semiconductor patterns are formed on the insulation film, it is well known that some wafers are electrified at a potential of several hundreds of volts. Under such high electrification, there is cause a problem that the auto-focus operation needs a relatively long time, resulting in the decrease in overall inspection throughput.
Further, when focusing is performed by the objective lens while the sample surface is electrically charged, the following problem is incurred. Whenever a charged particle beam system is used, the magnification is calculated on the assumption that the sample is not electrified at all. If the proper focus is not attained, the magnification is calculated on the assumption that the height of the sample has been changed. In case of a scanning electron microscope, however, if the sample surface is negatively charged, the electron beam is decelerated by the negative charge so that the focal point is shifted. If the shift of the focal point is adjusted by adjusting the exciting current for the objective lens and if the magnification is calculated depending on the adjusted exciting current, then the calculated magnification is different from the true magnification. In the case where the charged particle beam system is used to measure the dimensions of the patterns, a problem is incurred that the right dimensions cannot be determined.
The following embodiment of this invention will be proposed in consideration of the above mentioned problem.
The principle for measuring the electric potential of the sample surface will be described in reference to
Under the condition that Vr=−V1, where the accelerating voltage equals the retarding voltage, the kinetic energy of the primary electrons barely reaching the sample is 0 eV (zero electron volt). In this case, since the primary electrons cannot generate secondary electrons at the sample surface, the output of the secondary electron detector 10 goes down.
In the case where Vr>−V1, the primary electrons of the beam emitted from the electron gun hit the sample surface so that secondary electrons are emitted from the sample surface and detected by the secondary electron detector 10. Since the kinetic energy of the primary electrons reaching the sample surface increases as the retarding voltage Vr is increased, the amount of the emitted secondary electrons increases as the retarding voltage Vr is shifted in the positive direction.
As understood from the above description, If Vr<−V1, the output of the secondary electron detector 10 remains constant irrespective of the magnitude of the retarding voltage Vr; if Vr=−V1, where there is no secondary electron emitted, the output of the secondary electron detector 10 goes down; and if Vr>−V1, where the energy of the primary electrons hitting the sample surface varies depending on the magnitude of the retarding voltage, the amount of the secondary electrons emitted from the sample surface varies depending on the energy of the primary electrons so that the output of the secondary electron detector 10 varies depending on the magnitude of the retarding voltage.
In the foregoing description, the case is explained where the sample surface is not previously electrified. Now, description is made of the case with reference to
The method of measuring the potential of the sample surface will be described in detail with reference to the flow chart shown in
By sweeping the retarding voltage, the surface potential Vsp is calculated on the basis of the average brightness of the images obtained corresponding to the swept retarding voltages (S07). Then, the retarding voltage is controlled in such a manner that the beam of the primary electrons can hit the sample surface with a desired energy of Vacc eV (S08). Here, in consideration of the accelerating voltage V1 for the primary electrons and the surface potential Vsp, the retarding voltage Vr must be given a value such that
Vr=−V1+Vacc−Vsp (1)
In the description given for
Furthermore, the program to execute the operation according to this flow chart can be incorporated in a scanning electron microscope or a charged particle beam system.
According to the conventional auto-focus mechanisms described above which serve to evaluate the sharpness of pattern, the edge enhancing treatment typically known as the differential treatment must be employed in the evaluation of the obtained images. In order for the edge enhancing treatment to be effective, the S/N ratio of the image must be greater than a certain minimum value so that the noise contained in the image may not be enhanced. To keep the S/N ratio above a certain level in a scanning electron microscope, the ratio is generally improved by using the frame accumulation technique or by reducing the noise through the smoothing filter before the differential treatment. These measures, however, result in a problem that the operation of auto-focusing is lengthened. On the other hand, according to this invention, the sample surface potential can be measured even where there is no pattern on the sample surface. It is only necessary to obtain the average brightness of the acquired images. Accordingly, the auto-focus technique of this invention is little affected by noise and does not need to resort to the frame accumulation technique or the noise reduction through the smoothing filter for improving the S/N ratio. As a result, the auto-focus mechanism according to this invention can operate more quickly than the conventional auto-focus mechanism wherein the sharpness of pattern must be evaluated.
With reference to
Another embodiment is described below with respect to the data points used to obtain such a quadratic function as mentioned above. When the primary electrons cannot reach the sample surface, the output of the secondary electron detector or the average brightness of images is about constant. This constant value is, for example, set as a threshold. The data points whose values are less than this threshold are used to determine the minimum value. Curve fitting using quadratic functions is performed on these data points whose values are less than the threshold. Hence, there is obtained with high precision that value of the retarding voltage for which the output of the secondary electron detector is minimized even where the retarding voltage is discretely changed.
There, however, is a case where only the acquisition of the value of the retarding voltage for which the output I of the secondary electron detector is minimized is not always enough for the purpose, depending on the difference in the material of sample. In the variable range of retarding voltage where the primary electrons are turned back, the output of the secondary electron detector remains constant irrespective of the value of the retarding voltage or the difference in the material of sample. On the contrary, in the variable range of the retarding voltage where the primary electrons can reach the sample surface, since the amount of secondary electrons emitted from the sample depends on the efficiency of secondary emission of electrons, the Vr−I characteristic curve changes as seen in
Vr=−V1+Vacc−Vsp+Voff (2)
In general, it is impossible to obtain a properly focused image even though the surface potential is determined by using the above described method and the retarding voltage is controlled according to the above expression (2). That is because while the sample surface potential due to the charges on the surface can be offset by controlling the retarding voltage, the focus deviation due to the difference in the height of the sample is yet to be offset. In order to properly focus the electron beam on the sample surface, the focal point must be controlled by, for example, changing the exciting current for the magnetic objective lens. If the sample stage for holding the sample is capable of fine movement in the vertical direction, the height of the sample may be finely controlled by moving the sample stage vertically.
In this embodiment, the retarding voltage Vr is swept, but the accelerating voltage V1 for the primary electrons or both the accelerating voltage V1 and the retarding voltage Vr may be swept. For the surface potential can be determined by calculating the sum of or the difference between, the accelerating voltage V1 for the primary electrons and the value of the retarding voltage Vr for which the output of the secondary electron detector is minimized.
In this case, the determined surface potential may be superposed on the voltage to be applied to the sample stage or the electrode for accelerating the primary electrons, or shared in superposition between the sample stage and the electrode.
The retarding voltage may be increased in the positive or negative direction. If the retarding voltage is increased in the positive direction starting at the value thereof for which no secondary electron is emitted from the sample, then such an advantage can be obtained as follows. The value of the retarding voltage Vr for which the output of the secondary electron detector is minimized has only to be known to determine the surface potential. Therefore, the measurement can be finished at the time when the output of the secondary electron detector becomes minimum or when the data enough to effect the above mentioned curve fitting have been collected. The primary electrons have not yet hit the sample surface up to this time and the obtained advantage is that the damage to the sample due to the irradiation thereof by the primary electrons can be prevented and that the electrification of the sample due to the irradiation thereof by the primary electrons can also be avoided. Further, if the variable range of the retarding voltage is limited to the above mentioned region, the measurement can be completed in a short time, resulting in an improvement on the throughput.
Semiconductor circuit patterns are usually integrated in the surface of a sample called a wafer. A critical dimension measurement SEM is widely used, for example, to measure the width of a conductive line of such a pattern. The basic structure of such a dimension-measuring SEM is also as shown in
In the ordinary measurement of semiconductor circuit patterns, there is a case where several hundreds of measurement points in a single wafer are subjected to a required measurement at the stage of trial fabrication. It often happens that even after mass production is started, several tens of points in a single wafer are subjected to a prescribed measurement. In such a case, the throughput is deteriorated if the surface potential measurement is performed at all the measurement points according to the method described above in Embodiment 1. The deterioration of the throughput can be prevented by actually measuring the surface potentials at several points in a wafer before the measurement of pattern dimensions and accordingly estimating the surface potentials at arbitrary measurement points on the basis of the actually measured surface potentials.
In most cases, the entire surface of the wafer is electrified, with the central part of the wafer having a higher potential and the peripheral part thereof having a lower potential, as shown in
Vsp(x, y)=a×r3+b×r2+c×r+d r=(x2+y2)1/2 (3)
, where x and y are the abscissa and the ordinate of the rectangular coordinate system with its origin located at the center of the wafer, and a, b, c and d are constant coefficients.
The coefficients a, b, c and d may be determined through the use of the least square method so that the resultant cubic function can best fit the curve that approximates the data obtained by the measurement. In this example, the curve fitting function is derived on the assumption that the potential distribution is symmetric with respect to the center of the wafer, but any other type of expression for approximation may be employed if it is appropriate in approximating the potential distribution over the wafer surface. Moreover, the surface potential at an arbitrary point on the wafer may be determined by the interpolation on the basis of the data collected at the plural measurement points.
The dimension-measuring SEM enables an unattended operation in measurement by operating in accordance with a series of measuring steps recorded in a file referred to generally as a recipe. The measuring operation according to the recipe will be described below with reference to
At the beginning of the pattern dimension measuring steps, the first measuring point is reached by the movement of the sample stage (S40). Thereafter, the surface potentials at the measuring points are determined by using the surface potential approximating function Vsp(x, y) obtained in the step S30, the retarding voltage is controlled by using the above expression (2), and the effect of the surface potential is offset at each measuring point (S41). After these steps, the auto-focus operation is performed (S42), the pattern to be measured is located (S43), and the dimensions of the targeted pattern are finally measured (S44).
In general, the amount of electrification of wafer surface may change largely depending on semiconductor manufacturing processes. It may also happens that different wafers have different amounts of surface electrification even in the same manufacturing process. According to the methods described in the foregoing, since the surface potentials at plural measuring points must be determined even when there is little or no electrification in the wafer surface, then the measurement efficiency cannot be said to be very high. The electrification of wafer is normally greater at the center of the wafer than at the periphery thereof. The surface potential is first measured in the wafer center to determine whether the electrification in the wafer center is considerable or not. If the electrification in the wafer center is considerable, the measurement of the wafer surface potential is performed, and if the electrification in the wafer center is negligible, such measurement is omitted. Accordingly, the deterioration of the throughput can be prevented.
The flow of the above described series of steps is illustrated in
The screen image illustrating how the wafer surface potential is measured according to the recipe will now be described with reference to
In the above described screen image, three independent pieces of data are inputted. In the application of the sample surface potential measuring technique according to this invention, as already described in Embodiment 1, there should not necessarily be patterns in the sample surface and therefore the surface potential measurement may be performed at any measuring point. Consequently, when a recipe is prepared, it is possible to considerably reduce the number of the pieces of data inputted by a user. For example, only the number of the measuring points may be defined while the measuring points where the surface potentials are actually measured are randomly selected by the system itself. Alternatively, it is only specified whether the surface potential offset is performed or not, and the surface potential measurement may be performed at the plural coordinate points recorded previously in the system.
According to the above described method, several to ten measuring points are enough for the surface potential measurement and therefore the deterioration of throughput can be limited to the minimum. Further, since the wafer surface potential can be appropriately controlled, the sweeping range for the auto-focus operation can be narrowed so that the time required for the auto-focus operation can be shortened. Even in case where the wafer surface potential is measured, if there are many measuring points for pattern dimension measurement, the auto-focus operation at each measuring point can be omitted or the time required therefore can be shortened. Consequently, the time required for executing the recipe can be shortened. Furthermore, since the surface potential is controlled, the error of magnification arising from the surface electrification can be eliminated and the correct measurement of dimensions can be secured.
It should be further understood by those skilled in the art that although the foregoing description has been made on embodiments of the invention, the invention is not limited thereto and various changes and modifications may be made without departing from the spirit of the invention and the scope of the appended claims.
Number | Date | Country | Kind |
---|---|---|---|
2006-079908 | Mar 2006 | JP | national |
Number | Name | Date | Kind |
---|---|---|---|
4912052 | Miyoshi et al. | Mar 1990 | A |
4922097 | Todokoro et al. | May 1990 | A |
6667476 | Todokoro et al. | Dec 2003 | B2 |
6946656 | Ezumi et al. | Sep 2005 | B2 |
7239148 | Suhara | Jul 2007 | B2 |
7514681 | Marella et al. | Apr 2009 | B1 |
20040183013 | Nakasuji et al. | Sep 2004 | A1 |
20060163477 | Nozoe et al. | Jul 2006 | A1 |
20090057557 | Cheng et al. | Mar 2009 | A1 |
Number | Date | Country |
---|---|---|
4-229541 | Aug 1992 | JP |
10-125271 | May 1998 | JP |
2001-52642 | Feb 2001 | JP |
2005-166542 | Jun 2005 | JP |
Number | Date | Country | |
---|---|---|---|
20070221845 A1 | Sep 2007 | US |