EP1508073A2

EP1508073A2 - Method for determining wavefront aberrations

Info

Publication number: EP1508073A2
Application number: EP03737967A
Authority: EP
Inventors: Wolfgang Emer; Paul Gräupner
Original assignee: Carl Zeiss SMT GmbH
Current assignee: Carl Zeiss SMT GmbH
Priority date: 2002-05-24
Filing date: 2003-05-20
Publication date: 2005-02-23
Also published as: US7209241B2; US7019846B2; AU2003245886A1; JP2005527117A; WO2003100525A3; US20060164655A1; DE10224363A1; US20040032579A1; AU2003245886A8; WO2003100525A2

Abstract

The invention relates to a method for determining wavefront aberrations for the characterisation of the projection features of an optical projection system. According to said method, the measurement results of two different measurement methods that are carried out in succession are combined. At least one part of the aberration parameters that have been determined in the preceding first measurement method is assumed as a given for the determination of aberration parameters using the second measurement method and is used accordingly. A hybrid method is thus provided, in which the strengths of at least two measurement methods are used in combination and specific weaknesses of the methods can thus be avoided.

Description

Description Method for the determination of wavefront aberrations

The invention relates to a method for determining wavefront aberrations for the characterization of

Imaging properties of an optical imaging system. The preferred field of application is the measurement of projection objectives for microlithography.

Microlithographic projection exposure systems are used for the production of semiconductor components and other finely structured components. Here, a pattern of a mask or a reticle is imaged on a substrate covered with a light-sensitive layer using a projection lens. The finer the structures to be imaged, the more the quality of the products produced is determined and limited by imaging errors in the optical imaging systems used. These aberrations have an influence, for example, on the line widths shown and the image position of the structures shown.

The imaging properties are usually characterized using wavefront aberrations in order to obtain a quantitative measure of the deviation of the real image from an ideal image. The determination of wavefront aberrations is a crucial step in the manufacturing process of optical imaging systems in order to be able to provide systems with minimal imaging errors through suitable adjustment. Since the imaging quality of high-performance optical systems also critically depends on environmental influences such as temperature, pressure, mechanical stress and the like, monitoring of the imaging quality is also carried out at the customer's location as well as, if necessary aberration control through manipulation of the imaging system is essential. For this purpose, reliable, sufficiently precise measurement methods must be available which allow the projection objectives to be measured quickly in situ, ie in the installed state in a wafer stepper or wafer scanner.

The characterization of aberrations by

Wavefront aberrations are based on the consideration that when imaging a punctiform object through an ideal lens, the spherical wave emanating from the object on the image side of the lens runs again as a spherical wave to the punctiform image which lies in the center of the image-side spherical wave. In the case of a real lens with aberrations, the shape of the wavefront on the image side will deviate from a spherical shape, so that the image-side light rays will not unite in a point-like image, but in a blurred image. In order to enable a quantitative description of the image errors generated, the wave front that intersects the exit pupil of the imaging system on the optical axis is usually considered. The distance (in nm) between the actual and ideal wavefront is called the wavefront aberration. The wavefront aberration function is generally complex in shape. This function is usually written as the sum of standard functions Z-. Various groups of functions of Zj can be used for the purpose of aberration characterization. The so-called "Zernike polynomials" are usually used in the field of microlithography. The Zernike polynomials or corresponding Zernike coefficients can be derived or extracted from various measurement methods.

There are several methods available for determining the wavefront aberrations on site, with between direct and indirect methods for determining the current state of a lens can be distinguished.

In a known direct method, hereinafter referred to as the LITEL method, a locally tilted wave front is converted into distortion in the image plane with the aid of a specially constructed reticle. This is then measured using a standard box-in-box method. The wavefront is then reconstructed from this. The method is sufficiently accurate for most applications. However, the analysis time is in the hour range.

With the available indirect methods, measurements are carried out in different NA and multiple illumination settings (MIS). A distinction can be made here between aerial image measurements and resist profile measurements. A resist-based measurement technique is the so-called aberration ring test (ART) (see, for example, the articles “Impact of high order aberrations on the performance of the aberration monitor” by P. Dirksen, C. Juffermans, A. Engelen, P. De Bisschop, H. Muellerke, Proc. SPIE 4000 (2000), pages 9ff or "Application of the aberration ring test (ARTEMIS ™) to determine lens quality and predict its lithographic performance", by M. Moers, H. van der Laan, M. Zellerrath, W. de Boeij, N. Beaudry, KD Cummings, A. van Zwol, A. Becht and R. Willekers, lecture at SPIE, the 26th Annual International Symposium on Microlithography, February 25 to March 2, 2001, Santa Clara CA.). In the aberration ring test, a ring-shaped object is formed in the image plane. The deformations that can be measured on the depicted object with regard to ring diameter and ring shape in a focus series can be analyzed in the form of Fourier components, each Fourier component corresponding to a specific class of lumped aberrations, for example spherical, coma, astigmatism and triple ripple. It will assume that there is an essentially linear relationship between the Fourier components that can be determined in this way and the wavefront aberrations that can be described by Zernike coefficients, so that individual Zernike coefficients can be extracted from the "lumped aberrations" on the basis of a suitable model. Disadvantage This method is the dependency of the accuracy on the quality of the model used in the simulation.A comparison with aberration data, which were obtained with the help of highly precise direct, interferometric measurements of wavefront aberrations (cf. DE 101 09 929), shows that the current models The method becomes more unstable the more orders of Zernike coefficients are to be separated.

The invention has for its object to provide a method for determining wavefront aberrations, which makes it possible to determine wavefront aberrations with high accuracy on the basis of measurements that can be carried out in a short amount of time on site.

To achieve this object, the invention provides a method with the features of claim 1. Preferred developments are specified in the dependent claims. The wording of all claims is incorporated by reference into the content of the description.

In the method according to the invention, the wavefront aberration caused by the imaging system is first determined with the aid of a first measurement method. On the basis of the first measurement method, at least one first aberration parameter is determined from the determined wavefront aberrations. Usually a set with several first aberration parameters is determined.

Aberration parameters in the sense of this application are suitable Numerical or functional representations that allow or represent a quantification of the wavefront aberrations determined by the measurement method. A second measuring method is used at different times from the first measuring method, which can be based on a different measuring principle than the first measuring method. With the help of this second measurement method, wavefront aberrations caused by the imaging system are also determined. Based on these wavefront aberrations, at least a second one

Aberration parameter determined, which characterizes the wavefront aberration according to the second measurement method. Certain properties of such coefficients can also be used, for example the field dependence of Zernike coefficients. According to the invention, at least one first aberration parameter is used when determining second aberration parameters. This means that the result of the first measurement process flows in via this at least one first aberration parameter used in the evaluation of the second measurement process. In this way, a hybrid method is created in which the strengths of at least two measuring methods can be used in combination without the specific weaknesses of the methods having to impair the measurement.

In order that the results of the various measurement methods used can be compared with one another and offset against one another, a “common language” at the level of the measurement results is advantageous. It is advantageous if the first measurement method is used to determine a first set of first Zernike coefficients, which characterize the wavefront aberration measured with the first measurement method and serve as first aberration parameters. The wavefront aberrations determined with the second measurement method can be represented with the aid of a second set of second Zernike coefficients. Suitable coefficients can be selected from the first Zernike coefficients and used in the determination the second Zernike Coefficients are taken into account. Here too, a set with several such aberration parameters is preferably determined. Other aberration parameters suitable for describing wavefront aberrations can also be used, in particular those that can be converted into Zernike coefficients.

The inventors have found that there are certain aberrations among the contributions to the overall wavefront aberration that are relatively insensitive to environmental influences such as temperature, pressure or mechanical influences. These are referred to below as stable, fluctuation-insensitive or non-fluctuation-prone. In contrast, other Zernike coefficients are relatively susceptible to failure. These are referred to as unstable, susceptible to fluctuation or sensitive to fluctuation. If one considers a power series approach for image errors and their field course for a rotationally symmetrical system, it can be derived from symmetry overlays that the wave aberration function depends on the three variables r ² , φ ² and r * φ, where r is the field radius, φ the pupil radius and r * φ is the scalar product of the two. Developing the wave aberration function into a power series of these three variables results in the terms shown in Table 1:

The inventors have found that interference of any kind on the system is initially reflected in the lower order. The scale and focus offset, ie terms of first order, often react most strongly to disruptive influences. Seidel aberrations follow in the next order. This discovery by the inventors was confirmed by sensitivity calculations on specific lens designs: Both with environmental influences, such as temperature and pressure, and with mechanical fluctuations in the lens spacing, the optical performance reacts in the order of the orders of the power series development. From this it can be deduced that higher order aberration contributions are more stable, ie relatively less susceptible to transport-related changes, changed environmental conditions, etc. than lower order aberration contributions. A large number of tests have shown that terms up to and including Seidel aberrations can be subject to significant, i.e. measurable, fluctuations, while the fluctuations in the higher order range are generally negligible and can be below the measurement accuracy of the methods used to determine them , With increasing measurement accuracy or increasing accuracy requirements, the boundary between fluctuation-sensitive terms (terms of lower orders) and relatively fluctuation-insensitive terms (higher orders) can be shifted to higher orders. On the basis of these considerations, it has proven to be advantageous if at least one stable, ie fluctuation-insensitive first aberration parameter is determined in the first measurement method, for example at least one Zernike coefficient of higher order and / or its field dependency, and if this has at least one stable aberration parameter is taken into account when determining or calculating the second aberration parameters when carrying out the second measuring method with its value or course determined in the first measuring method. When evaluating the second measurement method, it is therefore not necessary to vary all the parameters that can be varied when adapting Zernike coefficients or equivalent parameters to a measurement result. Rather, some values known from the previous measurement method can be assumed as given, so that only the other values from the current measurement are to be determined. As a result, the second measurement method can be evaluated much more stably and precisely.

In practice, the first measurement method is normally carried out at the place where the optical imaging system is manufactured and the second measurement method at the place of use of the optical imaging system, for example at the semiconductor chip manufacturer. At the manufacturer, wavefront aberrations are normally determined with the greatest accuracy in order to qualify the manufactured product. These can be recorded by a first set of first Zernike coefficients and / or their field dependency. This is followed by transport from the place of manufacture to the place of use. The second measuring method can then be carried out on site on the ready-to-use imaging system before the start of production or during production breaks. Since, according to the inventors, there is a difference between the delivery of an imaging lens to the manufacturer and the commissioning of a wafer scanner or wafer stepper at Customers only significantly change aberration terms of lower orders, it is unnecessary to measure the higher orders at the customer again. It is sufficient to determine the aberration terms of the lower fluctuation-prone orders using a suitable method. If the terms of higher orders are assumed to be known, the corresponding evaluation methods become more stable and precise. In addition, methods that can be carried out particularly quickly can be selected for the measurement at the site of use, which allow the fluctuation-prone terms to be determined with sufficient accuracy.

To carry out the first measurement method, all methods can be used which ensure sufficient measurement accuracy, in particular for aberrations of higher orders. Direct methods of the type mentioned at the outset are particularly suitable. Because of the high measurement accuracies that can be achieved, interferometric methods are preferred. Such a method with shearing interferometry is described in DE 101 09 929. The relevant disclosure content of this patent application is made by reference to the content of this description.

The second measurement method, which is normally to be carried out at the location of the imaging system, can be selected with regard to the required measurement duration and the measurement accuracy for the fluctuation-sensitive aberration terms to be determined. For example, any indirect method of the type mentioned at the beginning (aerial image measurements or resist profile measurements with different NA / lighting settings) can be used. A suitable method with aerial image measurement, in which an aerial image sensor in the area of the image plane of the imaging system scans the best focus positions in the x, y and z directions for different field points and for different lighting settings, is shown in the article "Aerial image measurement method for fast aberration setup and illumination pupil verification" by H. van der Laan, M. Dierichs, H. van Greevenbroek, E. McCoo, F. Stoffels, R. Pongers and R. Willekers, Proc. SPIE 4346 (2001), pages 394 - 407. The disclosure of this article is incorporated by reference into the content of this description.

The above and further features emerge from the claims and also from the description and the drawings, the individual features being realized individually or in groups in the form of sub-combinations in one embodiment of the invention and in other fields, and advantageous as well protective designs can represent.

1 shows an example for the correction of interferometry data with the measurement data of an in-situ aerial image measurement technique.

The invention is explained on the basis of an exemplary embodiment in which an interferometric measurement of wave aberrations is carried out with a shearing interferometer according to German patent application DE 101 09 929.0 at the manufacturer of a projection objective. The description of the measuring device there is made by reference to the content of this description. Any suitable method known to those skilled in the art can be used to determine aberration parameters that describe the measured wavefront aberrations. The described method enables the direct measurement of wave aberrations in the pupil of the microlithographic projection objective with an accuracy that is currently not achieved by any other measuring device. After a final acceptance of a manufactured projection lens, information about the wavefront aberrations generated by the lens is available in the form of a set of interferometry data, which are for example include the Zernike coefficients of lower and higher orders (for example up to Z35) and the field profiles of these Zernike coefficients. This data forms the first aberration parameters.

After final acceptance, the projection lens is packed securely and transported to a distant customer using suitable means of transport. At the customer, the projection lens is installed in a wafer stepper which, in addition to the projection lens, also contains a suitable lighting device and manipulation devices for reticles and wafers. The structure of such microlithographic projection exposure systems is known to the person skilled in the art. The wafer stepper can be delivered to an end customer, usually a manufacturer of semiconductor chips or other finely structured components.

With the built-in projection lens

Imaging properties determined using a suitable in-situ measurement technique, with test structures being mapped at different lighting and aperture settings in the example. The results of these images can be characterized by various parameters, for example the best focus position of the imaged structure in the x, y and z directions and / or the curvature and / or asymmetry of an aerial image (aerial image).

It is possible to derive these parameters, which represent the result of the measurements, from the wave aberrations of the lens. In the example, the so-called "linear model" is used for this, in which a parameter corresponding to the lighting and aperture setting is represented as the sum of weighted Zernike coefficients according to: Parameter "= ^ f • Z _} ,

0)

where s is the corresponding lighting and aperture setting, ^{Jn is} a weighting factor for the Zernike coefficient Z _n for lighting / aperture setting s and Z _{n is} the Zernike coefficient number n. It has been shown that this model is valid with sufficient accuracy for small aberrations, as are often present in younger projection lenses. The weighting factors are determined by theoretical simulations.

A combination of the mapping of test structures at different illumination / aperture settings with the linear model mentioned makes it possible to use an inverse method to draw conclusions from the measured lithographic parameters on wavefront aberrations in the pupil. However, only a few low-order Zernike coefficients can be determined in this way, and the accuracy is also limited (cf. article H. van der Laan et. Al. Proc. SPIE 4346 (2001), pages 394 to 407).

In-depth investigations and analyzes by the inventors, which are based, for example, on correlation of aberration measurement data at the lens manufacturer, at the manufacturer of the wafer stepper and at the end customer, or on the comparison of interferometry measurements at lenses that come back to the manufacturer for repair purposes, have shown that certain Zernike Coefficients are relatively constant over time, ie essentially stable or insensitive to fluctuations. On the other hand, groups of imaging errors can be defined in which a temporal variation is possible, for example due to transport influences and other environmental influences. In the example case, the following are Aberration parameters have been identified as unstable or susceptible to fluctuation:

- tunable lowest order astigmatism (square field course) - centerable lowest order astigmatism (linear field course)

- tunable lowest order coma (linear field pattern)

- Centerable coma of the lowest order (constant in the field)

- tunable lowest order spherical aberration (constant in the field).

In the case of all other imaging errors, it is assumed in this case that the interferometry data measured by the lens manufacturer or the imaging errors represented by them remain constant over time.

In a first step, these interferometric data are used to correct systematic errors that can occur in the above-mentioned wavefront construction from lithographic parameters using the linear model. The reason for these systematic errors is the influence of higher-order Zernike coefficients, which cannot be taken into account in the inverse method for wave front reconstruction. A suitable correction procedure can proceed as follows.

After an aerial photo measurement of aerial photo shifts in x-, y-

Direction (plane perpendicular to the optical axis) for eight different lighting / aperture setting settings, the Zernike coefficients can be calculated based on the linear model as follows: dx meas _ie n dx high _eΛ dx meas _κl2 dx hιgh _et2 dx meas _iet3 dx ^hi s S.sel3, Z2 Sχel3, ZU dx meas _κt4 dx high _et4 Sχet4, Z7 Sχel4, ZU dx meas _{i (! L5} dx high _{se (5} (2)

Sχet5, Z2 S.xel5, Zl S.xet5, ZU dx meas _set6 dx high _wl6 "14.

Sχel6, Z2 Sχet6, Zl Sχet6, ZU dx meas _iet1 dx igh _et7 Sχet7, Z2 Ssea, zι seil, ZU dx meas _selS dx highes S &&, Z2 Rs _.-, Z7 R ^* e / 8, Z14 _

Here, the vector dx high contains the sum of all pattern shifts in the x direction due to Zernike coefficients except for Z2, Z7 and Z14. A standard procedure for calculating Z2, Z7 and Z14 is based on equation (2) in the sense of a least-square fit. This can be written as follows when using a pseudo inverse of the sensitivity matrix:

Using the linearity of equation (3), the result Z can be separated into two parts:

Z = Z _Q - AAZZ (4)

The value Z ₀ is the result of the aerial photo measurement if dx high is neglected. The value ΔZ thus represents the influence of the Zernike coefficients except Z2, Z7 and Z14. This missing factor is independent of the Zernike coefficients Z2, Z7 and Z14.

On this basis, the correction procedure can be carried out easily. First, ΔZ is calculated once for each lens using equation (5).

The vector dx high is then calculated for all settings from the interferometrically determined Zernike coefficients Z _n using the linear model as follows:

dx high ∑ / i ^S '-' _* ietx. Zn • z "(6)" = 10.19.23.26.39.34

Subsequently, ΔZ is subtracted from the resulting Zernike coefficients for each aerial photo measurement.

In a second step, interferometry data is corrected using measurement data from in-situ measurements of the wavefront aberrations in the scanner. As mentioned, a description of the wavefront aberration with maximum accuracy is possible with the lens data measured by the lens manufacturer. As mentioned, it is assumed that the wave aberrations of the lens have changed after installation in the scanner. It is assumed here that only a certain number of wave aberrations or aberration parameters that describe this aberration have changed (unstable aberration parameters).

If one looks at the Zernike coefficients in their scanner-integrated form (this form is referred to as fingerprint or finger print in the following), it is assumed in the example that only the following parameters change: Z7 (tilt), Z7 (offset), Z8 (offset) and Z9 (offset). In the present example, exactly these parameters are now measured using a suitable in-situ measurement technique, i.e. Measured again with lithographic parameters, in particular using aerial imaging technology for wavefront reconstruction. A combined data set is then created from the result of this measurement and the interferometric data. In this way, the aberration parameters determined by the manufacturer are used for the measurement on site at the end customer and are incorporated into the measurement result there.

First, the field profiles of all Zernike coefficients are taken from the interferometric data.

Then these field profiles are manipulated in such a way that the parameters selected above (in the example Z7 (tilt) and Z7 / 8/9 (offset)) of the manipulated interferometry data correspond to the parameters measured with the in-situ measurement technology. The field profiles of all other Zernike coefficients remain unchanged. For a more detailed explanation, FIG. 1 shows an example in which interferometric data are combined with data from aerial image measurements. The three diagrams shown side by side each show the field profiles of Z7 (coma x), Z8 (coma y) and Z9 (spherical aberration) measured along the x direction at the wafer level (in the image plane of the projection objective). The solid lines marked with circles represent the field course according to the interferometric data determined by the manufacturer. The dotted lines marked with crosses represent the corresponding field profiles from the aerial photo measurement. With

Dashed lines marked with squares each show the combined result, which includes measurement data from the interferometric measurement and data from the aerial image measurement. This example shows that the interferometric data are not replaced, but only by "tilting and shifting" to the

Results of the in-situ measurement data can be adjusted. The "finger prints" of all other coefficients remain unchanged.

Claims

claims

1. A method for determining wavefront aberrations for the characterization of imaging properties of an optical imaging system with the following steps:

Determination of wavefront aberrations caused by the imaging system with the aid of a first measuring method; Determining at least one first aberration parameter for the wavefront aberration based on the first measurement method; Determination of those caused by the imaging system

Wavefront aberrations with the help of a second measurement method that is applied at different times from the first measurement method; Determination of at least one second aberration parameter from the wavefront aberration on the basis of the second measuring method using at least one first resulting from the first measuring method

Aberrationsparameters.

2. The method of claim 1, wherein with the first measurement method, a first set of first Zernike coefficients and with the second measurement method, a second set of second Zernike coefficients for characterizing the

Wavefront aberration is determined.

3. The method according to claim 1 or 2, in which in the first measuring method at least one substantially stable fluctuation first aberration parameter is determined, this at least one stable aberration parameter when determining the second aberration parameters when carrying out the second measuring method with its first

Measurement method determined value is taken into account.

4. The method of claim 3, wherein at least one stable first aberration parameter corresponds to a Zernike coefficient with an order greater than two.

5. The method according to any one of the preceding claims, wherein the first measurement method is carried out at the location of the production of the optical imaging system and the second measurement method is carried out at a location of the imaging method.

6. The method as claimed in one of the preceding claims, in which, between the implementation of the first measurement method and the implementation of the second measurement method, the imaging system is transported from a first measurement location to a second measurement location.

7. The method according to any one of the preceding claims, wherein the first measurement method is a direct method for the direct determination of wavefront aberrations.

8. The method according to any one of the preceding claims, wherein the first measurement method is an interferometric measurement method, in particular in the manner of shearing interferometry.

9. The method according to any one of the preceding claims, wherein the second measurement method is an indirect method for indirect

Determination of wavefront aberrations.

10. The method of claim 9, wherein the second measurement method comprises measuring an aerial image generated by the imaging system.

1. The method according to claim 9, wherein the second measurement method is a resist-based method, which at least one

Exposure of one coated with a light-sensitive layer

Includes substrate and a measurement of the exposed substrate.