DE102004022596A1 - Process to detect positional errors in photolithographic projections in semiconductor wafers with circuit patterns uses overlay measuring apparatus and projection simulation - Google Patents

Abstract

Process to detect positional errors in photolithographic projection, especially for semiconductor wafer circuit patterns, comprises using an overlay measuring apparatus to evaluate a set of values and find and correct systematic positional errors in each field using a projection simulation model. The set values are compared with the specification. Independent claims are also included for the following: (A) an overlay measuring apparatus; (B) a process to evaluate an overlay;and (C) a process for a memory as above.

Description

The The invention relates to a method for determining positional errors in the photolithographic projection, in particular in structuring a semiconductor wafer having a circuit pattern. The invention concerns about it In addition, a method for evaluating measurement results of an overlay measuring device and a Overlay measurement device which is suitable to carry out this method.

to Integrated circuit fabrication is commonly done on semiconductor wafers provided with different electrical properties layers applied and each lithographically structured. A lithographic Structuring step may consist of a photosensitive Apply Resist, this with a desired structure for the relevant Level to illuminate and develop, and then the thus resulting resist mask in the underlying layer in an etching step transferred to.

With the ever increasing integration densities of integrated circuits increase also the requirements for the positional accuracy of a on the semiconductor substrate to be projected structure. Especially if already pre-levels in underlying layers, eg. B. transmitted in a lithographic projection step Need to become ever stricter tolerance limits with respect to mutual alignment the structure currently being projected onto the substrate relative considered to the structures of the aforementioned levels who the to the functionality to ensure the circuit.

density Line-column patterns, such as those in the field of manufacture of dynamic random access memories (DRAM) are formed, have, for example, in the area of the first circuit levels line widths of 70, 90 or 110 nm. In modern technologies of DRAM production will require the accuracy needed to align two structures, which is also referred to as an overlay budget, due to the smaller expectant structure resolutions continue to sink. This is for example the tolerable positional inaccuracy in the 100nm process line only still about 20 nm. Current and future Process lines are thus sensitive to errors in positional accuracy.

For the lithographic Projection step of such a circuit pattern is usually for the first critical levels a wafer scanner uses that in comparison to a wafer stepper a higher one Resolving power. In a wafer scanner, the exposure of the photosensitive takes place Resists along an exposure slot. The semiconductor wafer is generally deposited on a substrate holder and for exposure moved to a corresponding position. Then that will be on one Mask arranged circuit pattern successively into individual exposure fields transferred to the photosensitive resist. It will be during the Exposure of an exposure field of the substrate holder and a Exposure slot defining aperture shifted against each other. The exposure slot passes over the exposure field due to the uniform movement of the substrate holder and the aperture. Usually the size is one Exposure field about 26 mm × 35 mm.

The Exposure of the individual exposure fields is usually carried out in such a way that the top of the semiconductor wafer into a pattern of exposure fields in the form of a matrix or a grid (English: grid) divided is successively and with the wafer scanner or the wafer stepper is exposed.

The Determination of the positional accuracy of two superimposed layers takes place during the Production of integrated circuits usually with so-called Overlay targets. These are two substructures that each separately mapped to each of the layers. The first Substructure can consist of a rectangular structural element, that of a frame-shaped second Substructure is surrounded. Overlay targets usually become arranged together with other alignment marks in the saw frame area. The structure described above is as a box-in-box brand or too known as a box-in-frame brand. Usually is the offset of the individual substructures to each other with a Overlay measurement device for example, an optical microscope, measured.

at The exposure of a semiconductor wafer with a wafer scanner are several Known effects that can lead to overlay errors. This overlay error can be general two categories of sources of error assign. For one thing Errors occur during exposure within an exposure field arise. These sources of error are usually considered intra-field errors or field error. On the other hand, there are sources of error known by dividing the semiconductor wafer into individual Exposure fields are caused and different for each exposure field could be. These sources of error usually become referred to as an inter-field error or grid error.

The infra-field errors include the following contributions. For example, the exposure in an exposure field may be rotated by a fixed angle or may be wrong by a constant magnification factor. In the first case one speaks of a rotation error, the second case is commonly referred to as magnification error. Rotation errors are caused, for example, by misadjustment of the reticle.

Inter field error when lighting an exposure field by the controller caused by the substrate holder. A source of error is through a Twist the position of the exposure field by one from the position of the exposure field on the semiconductor wafer dependent angle given (so-called grid rotation error). Another source of error This also depends on the position of the exposure field the semiconductor wafer dependent Scaling error (grid magnification error). Furthermore plays with the exposure with a wafer scanner nor the so-called translation error a role created by different scanning directions and different speeds in different scan directions of the wafer scanner. The translation error is usually smaller in magnitude than the other above mentioned mistakes.

The Intrafield and interfield errors can each be separated for a component in a first direction (eg the x-direction) and for a component indicate perpendicular to this (y-direction). To the exact contributions of the Intrafeld and Interfeld errors to be able to determine usually becomes one Simulation model of the projection apparatus used, for example for one Wafer scanner a calculation rule with ten parameters for correction the individual sources of error in subsequent exposure steps includes.

at The fabrication of integrated circuits typically involves multiple overlay targets for each Exposure field provided. However, it will not be for everyone Semiconductor wafers all Survey overlay targets, usually only a random examination of selected overlay targets takes place. To estimate to be able to Whether the present semiconductor wafer meets the overlay specification will become common made a statement for the entire semiconductor wafer from the sampling measurement. If the sample does not meet certain statistical criteria, it will the semiconductor wafer fed to a post-processing. This is usually the already textured resist layer is removed again, it will applied a new resist layer and then again structured by photolithography.

In a method known to the art is in a first step From the measured offset values for each overlay target an average value is determined. Subsequently the standard deviation is calculated from the measured offset values, from which a maximum overlay error is calculated, for example by the mean and the added triple value of Standard deviation is determined. Should this maximum overlay error above a predetermined threshold, the semiconductor wafer a post-processing supplied. However, this procedure is very inaccurate because it is assumed that the offset values obey a normal distribution.

Out DE-A1-100 48 809 a method is known, in particular a better determination of the largest positional error enable should, since it is no longer assumed that a normal distribution. According to this Reference will be the existing position errors in a parametric and a random one Split distribution. Subsequently an algorithm is used to determine the greatest positional error which make it possible should, the expected probability for the required minimum specification to be determined as well as under consideration the efficiency of the semiconductor wafer to be reworked.

The However, the above methods are associated with problems. The The former method has the particular disadvantage that systematic error components (eg when painting or when Measurements are not taken into account). At closer Contemplation leads the method according to DE-A1-100 48 809 to an underestimation of the systematic error components due to overestimated values.

It is therefore an object of the invention, in a first aspect, a method for determining positional errors in the photolithographic projection, in particular in the structuring of a semiconductor wafer with a circuit pattern, and, in a second aspect, a method of evaluation of measurement results of an overlay meter that the above overcome these problems.

• – Bereitstellen eines Halbleiterwafers mit einer auf einer Vorderseite des Halbleiterwafers aufgebrachten Resistschicht;
• – Bereitstellen von wenigstens zwei Schaltungsebenen zur Projektion je eines Schaltungsmusters zur Bildung einer integrierten Schaltung, wobei für die erste Schaltungsebene mehrere erste Messmarken vorgesehen sind, die geeignet sind, zusammen mit zweiten Messmarken der zweiten Schaltungsebene jeweils ein Overlay-Target zu bilden;
• – Bereitstellen eines Projektionsapparates;
• – Bereitstellen eines eine Berechnungsvorschrift enthaltenden Simulationsmodells für den Projektionsapparat;
• – Aufteilen der Vorderseite des Halbleiterwafers in eine Mehrfachanordnung von Belichtungsfeldern;
• – Sukzessives Projizieren des Schaltungsmusters in Belichtungsfelder auf die Resistschicht, um eine Resiststruktur zu bilden;
• – Aufteilen der Belichtungsfelder in eine erste Teilmenge und in eine zweite Teilmenge;
• – Bestimmen von ersten Versatzwerten für jedes Overlay-Target in jedem Belichtungsfeld der ersten Teilmenge, wobei sich die ersten Versatzwerte aus ersten systematischen Lagefehlern und einem mittleren Lagefehler zusammensetzen;
• – Bestimmen der systematischen Lagefehler in jedem Belichtungsfeld der ersten Teilmenge anhand eines Simulationsmodells des Projektionsapparats;
• – Bestimmen eines mittleren Lagefehlers anhand der um die systematischen Lagefehler korrigierten ersten Versatzwerte der Belichtungsfelder der ersten Teilmenge;
• – Bestimmen eines zweiten Versatzwertes für jedes Belichtungsfeld der zweiten Teilmenge, der sich aus zweiten systematischen Lagefehlern und dem mittleren Lagefehler zusammensetzt, wobei der Beitrag der zweiten systematischen Lagefehler anhand des Simulationsmodells für die Belichtungsfelder der zweiten Teilmenge bestimmt wird; und
• – Bestimmen einer Häufigkeitsverteilung der ersten Versatzwerte und der zweiten Versatzwerte, um zu bestimmen, welcher Anteil der Belichtungsfelder innerhalb eines vorgegebenen Versatzwerts liegt.

This object is achieved in its first aspect according to the invention in that the following steps are carried out:

• Providing a semiconductor wafer having a resist layer deposited on a front side of the semiconductor wafer;
• Providing at least two circuit levels for projecting a respective circuit pattern to form an integrated circuit, wherein for the first circuit plane a plurality of first measurement marks are provided which are suitable for forming an overlay target together with second measurement marks of the second circuit level;
• - Providing a projection apparatus;
• - Providing a calculation model containing a simulation model for the Projekti onsapparat;
• - dividing the front side of the semiconductor wafer into a multiple arrangement of exposure fields;
• Successively projecting the circuit pattern into exposure fields on the resist layer to form a resist pattern;
• Dividing the exposure fields into a first subset and a second subset;
• Determining first offset values for each overlay target in each exposure field of the first subset, the first offset values being composed of first bias errors and a mean pitch error;
• Determining the systematic position errors in each exposure field of the first subset on the basis of a simulation model of the projection apparatus;
• Determining a mean positional error based on the first offset values of the exposure fields of the first subset corrected by the systematic position errors;
• Determining a second offset value for each exposure field of the second subset, which is composed of second systematic position errors and the mean position error, wherein the contribution of the second systematic position errors is determined on the basis of the simulation model for the exposure fields of the second subset; and
• Determining a frequency distribution of the first offset values and the second offset values to determine which portion of the exposure fields is within a predetermined offset value.

One essential aspect of the invention in its first aspect It is photolithographically a semiconductor wafer with a circuit pattern to structure an integrated circuit and then a evaluate first subset of overlay targets to determine offset values. In this case, a first subset of first offset values is input systematic positional error in each exposure field of the first subset based a simulation model of the projection apparatus and a medium Position error based on the corrected by the systematic position errors determined first offset values of the exposure fields of the first subset. Starting from the simulation model of the projection apparatus second offset value for each exposure field of the second subset is calculated from second systematic position errors and the mean positional error put together. Thus, the exposure fields of the second Subset included in an estimation about whether the offset values meet a specific overlay specification. there become the systematic contributions supplied by the simulation model of the projection device, the statistical posts follow the measured values of the first subset. This leads to a accurate and consistent determination of the second offset values, without that certain assumptions regarding the statistical distribution or averaging the result impair could.

In a preferred embodiment The step of providing the circuit pattern comprises the circuit pattern of a substantially rectangular saw frame is surrounded and for the first circuit level provided at least four first measurement marks which are suitable, together with at least four second marks each of the second circuit level to form an overlay target, where the at least four overlay targets are each on one corner of the saw frame are arranged.

According to this Approach will be for each Exposure field provided at least four overlay targets, which are positioned at the extreme points of the exposure field. This allows a precise Determination of the offset values, as possible position errors are determined by the determine the entire dimension of the exposure field most accurately.

In a further preferred embodiment the step of providing the projection apparatus comprises that for the projection apparatus, a wafer scanner is provided.

According to this Approach is for the projection apparatus uses a wafer scanner, in particular for critical Circuit levels due to the resolution is preferred.

In a further preferred embodiment The step of providing the simulation model comprises provided a simulation model with at least ten parameters which is suitable, a translation error, a grid rotation error, a grid magnification error, a field rotation error and a field magnification error, respectively, in two specify perpendicular coordinate axes.

According to this The procedure will be the subsequent exposure steps correctable positional error in the projection of a circuit plane in Considered simulation model, wherein to improve the assignment of the positional error in each case two separate and mutually perpendicular coordinate axes be specified.

• – Bereitstellen eines Overlay-Messgeräts; und
• – Vermessen der Overlay-Targets in jedem Belichtungsfeld der ersten Teilmenge mit dem Overlay-Messgerät.

In a further preferred embodiment, the step of determining first offset values for each overlay target in each exposure field comprises the first subset

• - providing an overlay meter; and
• - Measuring the overlay targets in each exposure field of the first subset with the overlay meter.

According to this The procedure will be an art-overlay overlay meter for determination the first offset values for each overlay target in each exposure field of the first subset used. This allows a simple and reliable determination of the first Offset values, where usually Applicable devices can be used in semiconductor manufacturing.

In a further preferred embodiment The first offset values are respectively in the x-direction and the y-direction certainly.

According to this The procedure is an assignment of the first offset values and the in the calculation specification parameterized positional error of the projection apparatus allows to be precise Determination, for example, to carry out the systematic positional error.

In a further preferred embodiment gets over it out for each overlay target in each exposure field of the first subset from the first offset values and the systematic position errors a residual error value for each overlay target in each exposure field of the first subset certainly.

According to this Approach is for each overlay target in each exposure field of the first subset from the measured first offset values, a residual error value that determines the posts the systematic positional error is corrected. In an easy way can calculate the residual error value by subtracting the measured first Offset values and the calculated systematic position error determined become.

In a further preferred embodiment gets over it addition, the standard deviation of the residual error value in x- and y direction, which are then used to determine the ten parameters the simulation model of the projection apparatus is used, by determining the minimum value of the standard deviation.

The Minimizing the residual error value is the central concern in determining the systematic positional error. By calculation the standard deviation of the residual error values can be determined set up a simple linear system of equations that is optimal Determination of the parameters of the simulation model of the projection apparatus is used. This allows a precise and effective determination the parameter of the simulation model.

In a further preferred embodiment The step of determining the mean positional error comprises the first offset values corrected for the systematic position errors the exposure fields of the first subset that the mean the residual error values in the x and y direction as a mean positional error is determined.

According to this The procedure will take into account all exposure fields mean positional error determined, with the deviations of the individual residual error values in x and y direction as a measure of the statistical Contribution of the first offset values can be used.

In a further preferred embodiment gets over it In addition, from the residual error values in the x direction and y direction, a value for the statistical Fluctuation of the mean positional error based on a probability distribution certainly.

According to this Approach, a probability distribution is determined which indicates how the residual error values in the x-direction and y-direction fluctuate around the mean positional error. This can be the statistical Contribution to the first offset values using the probability distribution be determined, if necessary determined from a measurement or empirical values.

In a further preferred embodiment the probability distribution is a normal distribution and for the Statistical fluctuation of the mean positional error becomes the standard deviation the residual error values are respectively determined in the x-direction and y-direction.

According to this Approach is it possible for the Probability distribution to use a normal distribution, since This often enough the fluctuation of the average positional error describes. this makes possible a simple determination of the statistical contribution to the offset values.

In a further preferred embodiment includes the step of dividing the exposure fields into the first subset and the second subset that the first subset approximately 10% of the exposure fields.

According to this procedure, only a small part of the exposure fields is measured, for example with an overlay measuring device, which leads to a higher throughput in the wafer inspection. However, due to the determination of the second offset values in the exposure fields of the second subset which are not evaluated by measurement, this does not lead to a deterioration of the quality control le.

In a further preferred embodiment The step of providing the circuit levels comprises a circuit design of a semiconductor memory with dynamic Memory cells comprising trench capacitors is provided, the smallest dimensions of 100 in the area of trench capacitors nm or less.

Semiconductor memory with dynamic storage cells are often in a high-volume manufacturing process produced and have a limited due to the small structures Overlay budget on. The inventive method speaks exactly this Problem area in which the determination of the overlay specification saves time based on a subset of measured overlay targets, in particular, the accuracy even at very tight specifications important is. This can be the production of semiconductor memory ei ne fast and accurate statement about the quality of Overlays are taken.

In a further preferred embodiment The step of determining the frequency distribution of the first comprises Offset values and the second offset values that the default Offset value approximately 20 nm.

According to this Approach is achieved that with modern technologies provided values of the overlay budget could be used to the semiconductor wafer according to the frequency distribution of the first Characterize offset values and the second offset values.

In a further preferred embodiment The step of determining the frequency distribution of the first comprises Offset values and the second offset values that the default Proportion of exposure fields within the specified offset value at least 99%.

at The manufacture of integrated circuits must have many different specifications be adhered to in order to achieve a high yield. To size Position inaccuracies could lead to failure of the circuit, is preferably a high Proportion of exposure fields within the specified offset value required.

• - Zuführen des Halbleiterwafers zu einer Nachbearbeitung.

In a further preferred embodiment, the following step is carried out for semiconductor wafers whose first offset values and second offset values have a lower fraction.

• - Feeding the semiconductor wafer to a post-processing.

Semiconductor wafer, have too many exposure fields with too large offset values, be according to this Procedure not further produced. In a postprocessing can be possibly correct the errors that occur, so that the already processed semiconductor wafers are not lost. This allows a cost-effective Production of integrated circuits.

• – Entfernen der Resistschicht;
• – Aufbringen einer weiteren Resistschicht auf die Vorderseite des Halbleiterwafers; und
• – erneutes Projizieren des Schaltungsmusters in die Belichtungsfelder auf der weiteren Resistschicht.

In a further preferred embodiment, the post-processing step comprises:

• - removing the resist layer;
• Applying a further resist layer to the front side of the semiconductor wafer; and
• Re-projecting the circuit pattern into the exposure fields on the further resist layer.

Prefers The measurement will be the wafer inspection before transferring the resist mask executed in the underlying layer. As a result, for semiconductor wafers, have too many exposure fields with too large offset values, the resist mask will be removed again and the last projection step will be repeated. This allows cost-effective production of integrated Circuits.

• – Bereitstellen eines Eingangsdatensatzes, der eine Vielzahl von ersten Versatzwerten umfasst, wobei die ersten Versatzwerte aus Messungen einer ersten Teilmenge von Overlay-Targets zweier Schaltungsebenen, die in einer Mehrfachanordnung von Belichtungsfeldern auf einem Halbleiterwafer mit einem Projektionsapparat aufgebracht wurden, bestimmt wurden und sich aus ersten systematischen Lagefehlern und einem mittleren Lagefehler zusammensetzen;
• – Bereitstellen eines eine Berechnungsvorschrift enthaltenden Simulationsmodells für den Projektionsapparat;
• – Bestimmen der systematischen Lagefehler in jedem Belichtungsfeld der ersten Teilmenge anhand des Simulationsmodells des Projektionsapparats;
• – Bestimmen eines mittleren Lagefehlers anhand der um die systematischen Lagefehler korrigierten ersten Versatzwerte der Belichtungsfelder der ersten Teilmenge;
• – Bestimmen eines zweiten Versatzwertes für jedes Belichtungsfeld einer von der ersten Teilmenge verschiedenen zweiten Teilmenge, der sich aus zweiten systematischen Lagefehlern und dem mittleren Lagefehler zusammensetzt, wobei der Beitrag der zweiten systematischen Lagefehler anhand des Simulationsmodells für die Belichtungsfelder der zweiten Teilmenge bestimmt wird; und
• – Bestimmen einer Häufigkeitsverteilung der ersten Versatzwerte und der zweiten Versatzwerte, um zu bestimmen, welcher Anteil der Belichtungsfelder innerhalb eines vorgegebenen Versatzwerts liegt.

The object is achieved in a further aspect according to the invention by a method for evaluating measurement results of an overlay measuring device, comprising the following steps:

• Providing an input data set comprising a multiplicity of first offset values, the first offset values being determined from measurements of a first subset of overlay targets of two circuit levels, which were applied in a multiple array of exposure fields on a semiconductor wafer with a projection apparatus, and from first systematic errors and a medium misalignment;
• - Providing a calculation model containing a simulation model for the projection apparatus;
• Determining the systematic position errors in each exposure field of the first subset based on the simulation model of the projection apparatus;
• Determining a mean positional error based on the first offset values of the exposure fields of the first subset corrected by the systematic position errors;
• Determining a second offset value for each exposure field of a second subset different from the first subset, which is composed of second systematic position errors and the mean position error, the contribution of the second systematic position errors being determined using the simulation model for the exposure fields of the second subset; and
• Determining a frequency distribution of the first offset values and the second offset values to determine which portion of the exposure fields is within a predetermined offset value.

One essential aspect of the invention in its second aspect is to evaluate data from an overlay meter such that from a data set, the first offset values measured by the overlay meter includes calculating second offset values representative of those exposure field positions be determined for that have no first offset values. It will be from the first Subset of first offset values a systematic misalignment in each exposure field of the first subset based on a simulation model of a Projection apparatus used in projection and middle one Positional errors are corrected on the basis of the systematic positional errors determined first offset values of the exposure fields of the first subset. Starting from the simulation model of the projection apparatus second offset value for each exposure field of the second subset is calculated from second systematic position errors and the mean positional error put together.

In a preferred embodiment, the following calculation is carried out to determine systematic positional errors on the basis of the simulation model: Ovl_X (x, y) = TransX + CF (x) * GMagX - CF (y) * GRotY + IF (x) * FMagX - IF (y) * FRotY, and Ovl_Y (x, y) = TransY + CF (y) * GMagY + CF (x) * GRotX + IF (y) * FMagY + IF (x) * FRotX; where x is the position of the exposure field along the first coordinate axis, y the position of the exposure field along the second coordinate axis, Ovl_X (x, y) or Ovl_Y (x, y) the systematic position error, TransX or TransY parameters of the translation error, GMagX or GMagY parameters of the grid magnification error, GRotX or GRotY parameters of the grid rotation error, FMagX or FMagY parameters of the field magnification error and FRotX or FRotY parameters of the field rotation error in the x-direction and y-direction respectively and CF or IF represent a decomposition of the coordinate values into a contribution on the semiconductor wafer or within an exposure field.

According to this Approach, a calculation rule can be used the ones usually dominant contributions the positional error in the projection parameterized with a wafer scanner. This allows a simple determination of the individual contributions, the example below for correction in further exposures of others Semiconductor wafer can be used.

In a further preferred embodiment, the residual error value is calculated according to the following equations: Res_X = M_X - Ovl_X_ (x, y) and Res_Y = M_Y - Ovl_Y_ (x, y), where Res_X or Res_Y represent the residual error value in the x or y direction and M_X or M_Y the first offset value in the x and y directions, respectively.

According to this The procedure becomes simple calculation steps, the measurement results of a Overlay targets around the systematic positional errors generated by the projection apparatus come, corrected. The remaining or residual error can occur This way can be easily determined.

• – Berechnen der Werte Ovl_X(x,y) bzw. Ovl_Y(x,y) der systematischen Lagefehler in x- und y-Richtung an wenigstens vier Positionen innerhalb des Belichtungsfeldes;
• – Bestimmen der maximalen Werte der systematischen Lagefehler jeweils in x- und y-Richtung;
• – Bestimmen des zweiten Versatzwertes in x-Richtung anhand folgender Gleichung: B_X = Ovl_X(x,y) + 3·SIGMA, wobei x die Position des Belichtungsfeldes, B_X den zweiten Versatzwert in x-Richtung, Ovl_X(x,y) den aus dem Simulationsmodell bestimmten systematischen Lagefehler in x- Richtung und SIGMA die Standardabweichung der residualen Fehlerwerte repräsentieren; und
• – Bestimmen des zweiten Versatzwertes in y-Richtung anhand folgender Gleichung: B_Y = Ovl_Y(x,y) + 3·SIGMA, wobei y die Position des Belichtungsfeldes, B_Y den zweiten Versatzwert in y-Richtung und Ovl_Y(x,y) den aus dem Simulationsmodell bestimmten systematischen Lagefehler in y-Richtung und SIGMA die Standardabweichung der residualen Fehlerwerte repräsentieren.

In another preferred embodiment, the step of determining the second offset values for each exposure field of the second subset comprises:

• Calculating the values Ovl_X (x, y) and Ovl_Y (x, y) of the systematic position errors in the x and y directions at at least four positions within the exposure field;
• Determining the maximum values of the systematic position errors in the x and y direction in each case;
• Determining the second offset value in the x direction using the following equation: B_X = Ovl_X (x, y) + 3 * SIGMA, where x is the position of the exposure field, B_X is the second offset value in the x direction, Ovl_X (x, y) is the the systematic error in the x-direction determined by the simulation model and SIGMA represent the standard deviation of the residual error values; and
• Determining the second offset value in the y direction using the following equation: B_Y = Ovl_Y (x, y) + 3 * SIGMA, where y is the position of the exposure field, B_Y is the second offset value in the y direction, and Ovl_Y (x, y) is the The systematic deviation in the y-direction and SIGMA, which represent the simulation model, represent the standard deviation of the residual error values.

According to this procedure, an offset value is calculated for each exposure field assigned to the second subset, taking into account the systematic contributions via the simulation model and the statistical contributions via the standard deviation of the residual error values of the first subset. This gives a precise prediction for all exposure fields those in which no measurement of offset values has been performed. Using the calculation rules of the simulation model, the systematic position errors in the x-direction and y-direction are taken into account with the contributions adapted to the actual conditions in the projection. As a result, the location dependence of the systematic position errors can be determined on the basis of the simulation model. The statistical contributions, in turn, are derived from the measured offsets, so that the offsets of the second subset are estimated with great accuracy and free from arbitrary assumptions.

• – Bereitstellen eines Datenfeldes mit ersten Speicherstellen, die geeignet sind, wenigstens einen Zahlenwert zu speichern, wobei jeder ersten Speicherstelle jeweils ein Bereich von Versatzwerten zugeordnet wird, so dass das Datenfeld Versatzwerte innerhalb einer vorgegebenen unteren Grenze und einer vorgegebenen oberen Grenze abdeckt; und
• – Zuweisen jedes ersten Versatzwertes und zweiten Versatzwertes an eine der ersten Speicherstellen, wobei der Inhalt der Speicherstelle die Anzahl der ihr zugewiesenen ersten Versatzwerte oder zweiten Versatzwerte umfasst.

In another preferred embodiment, the step of determining the frequency distribution of the first offset values and the second offset values comprises:

• Providing a data field having first memory locations adapted to store at least one numerical value, wherein each first memory location is assigned a range of offset values such that the data field covers offset values within a predetermined lower limit and a predetermined upper limit; and
• Assigning each first offset value and second offset value to one of the first memory locations, wherein the content of the memory location comprises the number of the first offset values or second offset values assigned to it.

According to this The procedure becomes the first offset values and second offset values stored in a histogram by the memory cells represents becomes. This allows in particular a simple determination of the frequency distribution the first offset values and the second offset values.

In a further preferred embodiment represent the lower limit has an offset value of about 5 nm and the upper limit an offset value of approximately 25 nm and represents the area of a memory location is about 0.25 nm.

According to this The procedure is the data field with the first memory cells divided into areas similar to those used in modern manufacturing processes Cover range of offset values. This allows the process, for example used in the manufacture of semiconductor memories.

In a further preferred embodiment In addition, the second memory cells are suitable, the number the first offset values or second offset values assigned to it each separately for a proportion in the direction of a first coordinate axis and a for one Store proportion in the direction of a second coordinate axis.

According to this The data field is suitable, the frequency distribution of the first offset values and the second offset values separated in x-direction and y-direction save. this makes possible one for both directions separate determination of whether a given overlay specification is complied with.

In a further preferred embodiment is a counter provided, which is suitable, the total number of first offset values and second offset values, and in which the content of the second memory cells is normalized to the total number.

According to this The procedure becomes a normalized frequency distribution of the first Offset values and the second offset values are determined, causing the Evaluation of the frequency distribution carried out in a simple manner can be.

advantageous Further developments of the invention are specified in the subclaims.

The The invention will now be described with reference to the accompanying drawings. In show the drawing:

1 an exposure apparatus in a schematic cross-sectional view of the application of the method according to the invention according to the first sub-aspect;

2 an overlay measuring device in a schematic cross-sectional view of the application of the method according to the invention according to the first and the second aspect;

3 in a plan view schematically the front of a semiconductor wafer in the application of the method according to the invention;

4 in a further plan view schematically the front of a semiconductor wafer in the application of the method according to the invention;

5 in a diagram, offset values which are determined during the application of the method according to the invention to the first and second partial aspect;

6 in a diagram, a cumulative and normalized representation of offset values, which are determined during the application of the method according to the first and second sub-aspects;

7 in a diagram an enlarged section of the diagram according to 6 ;

8th an embodiment of the method according to the first aspect in a flowchart; and

9 an embodiment of the method according to the second aspect in a flowchart.

The inventive method will be below for an exposure in an exposure field of a first level with Described a wafer scanner and a second level with a wafer stepper each the same size of the exposure field exhibit. The invention is also applicable to several circuit levels applicable with different configuration of the exposure devices used. So it is also possible several re the smallest dimensions critical circuit levels with a wafer scanner to expose and then to perform an exposure with a wafer stepper.

In 1 is a schematic cross-sectional view of the structure of a projection apparatus 5 shown. The projection apparatus 5 includes a movable substrate holder 12 , On the substrate holder 12 is a semiconductor wafer 10 filed on the front of a resist layer 14 is applied for example by spin coating.

The projection apparatus 5 further includes a light source 16 that over the substrate holder 12 is arranged, and is suitable to emit light, for example, with a wavelength of 248 nm, 193 nm or 157 nm. That from the light source 16 radiated light is transmitted through a projection lens 20 on the surface of the semiconductor wafer 10 projected.

Between the light source 16 and the projection lens 20 is a reticle 18 attached, which is provided with a pattern of a circuit level. For a wafer scanner, there is an exposure slot between the reticle 18 and the projection lens 20 attached (not in 1 shown). By controlling the substrate holder 12 becomes the front side of the semiconductor wafer 10 successively structured in individual exposure fields.

For the circuit levels For example, a circuit design of a semiconductor memory with dynamic memory cells comprising trench capacitors, the smallest in the range of trench capacitors of 100 nm or less.

After the photolithographic projection, the positional accuracy of the currently exposed layer relative to existing levels with an overlay meter 22 certainly. For this purpose, the first circuit levels already transferred into a layer or the substrate have been provided with measuring marks which are suitable for forming an overlay target together with measuring marks of the currently projected second circuit level. The overlay meter 22 includes a further substrate holder 23 which is suitable for the semiconductor wafer 10 take.

On the front of the semiconductor wafer 10 is a resist structure 14 ' applied by developing from the resist layer 14 originated is. The overlay meter 22 also includes another light source 24 to the front of the semiconductor wafer 10 to be irradiated with light. That from the front of the semiconductor wafer 10 reflected light is in a microscope 26 detected, with a processing means, such as a processor 27 connected is. By means of the microscope 26 and the processor 27 Measurement data are processed by overlay targets of the individual exposure fields.

This is where the inventive method in its two aspects. The exposure fields are assigned to a first subset and a second subset, as in 3 is shown. Each exposure field has a plurality of overlay targets, which are from first measurement marks in the first circuit level and second measurement marks in the second circuit level. Exemplary is in 3 an exposure field 28 shown that five overlay targets 30 includes, wherein at least four overlay targets 30 are arranged on a respective corner of a frame surrounding the exposure field.

The division of the exposure fields into the first subset and the second subset is carried out so that the first subset contains only a small portion, for example 10%, of the exposure fields. This is the time required for the measurement of the semiconductor wafer 10 with the overlay meter 22 significantly reduced, resulting in a higher throughput.

The first measuring marks 32 the first circuit level are in 3 by circular symbols and second measuring marks 34 the second circuit level shown in the form of a cross. In 3 For example, a total of nine exposure fields are shown, each comprising five overlay targets and together the first subset 36 represent. By measuring these overlay targets, one obtains the first offset values of the first subset 36 ,

As already mentioned, the first offset values were set for each overlay target 30 the first subset of a systematic position errors and a mean positional error together. When determining the first offset values by means of the overlay measuring device, the first offset values are each separated for a first coordinate axis, which in 3 is denoted as the x-direction, and for a second coordinate axis in 3 as y-direction is determined.

The following is the systematic registration error in each exposure field 26 the first subset 36 on the basis of a simulation model of the projection apparatus 5 certainly. This will be a simulation model of the projection apparatus 5 which is adapted to indicate a translation error, a grid rotation error, a grid magnification error, a field rotation error, and a field magnification error, respectively, in two orthogonal coordinate axes.

In order to determine systematic positional errors with the simulation model, the following calculations are carried out, for example, for a CANON wafer scanner: Ovl_X (x, y) = TransX + CF (x) * GMagX - CF (y) * GRotY + IF (x) * FMagX - IF (y) * FrotY, and Ovl_Y (x, y) = TransY + CF (y) * GMagY + CF (x) * GRotX + IF (y) * FMagY + IF (x) * FrotX.

at this calculation specification x denotes the position of the exposure field along the first coordinate axis and y the position of the exposure field along the second coordinate axis. The aim of the calculation is the systematic positional error Ovl_X (x, y) or Ovl_Y (x, y) in both To determine coordinate directions. This positional error is from the following contributions together: TransX or TransY are parameters of the translation error, GMagX or GMagY represent Parameters of the grid magnification error, GRotX or GRotY parameters of the grid rotation error, FMagX resp. FMagY parameters of the field magnification error and FRotX or FRotY parameters of field rotation error. These are given in x-direction and y-direction respectively.

The symbols CF and IF represent a decomposition of the coordinate values into a contribution on the semiconductor wafer or within an exposure field, ie for the coordinate values in the x-direction and y-direction, the following applies: x = CF (x) + IF (x), and y = CF (y) + IF (y).

Subsequently, will for each Overlay target in each exposure field of the first subset of the first offset values and the systematic positional errors a residual Error value is determined.

The residual error value is calculated for each overlay target in each exposure field of the first subset according to the following equations: Res_X = M_X - Ovl_X (x, y), and Res_Y = M_Y - Ovl_Y (x, y).

there represent Res_X or Res_Y the residual error value in the x or y direction and M_X or M_Y the first offset value in the x and y directions, respectively.

Subsequently, will the standard deviation of the residual error value in the x and y direction certainly. Their value is calculated from the root of the sum of Squares the residual error value in the x or y direction. These Calculation is known in the art and therefore will not be here explained in more detail.

Subsequently, will the standard deviation of the residual error value in the x and y direction for determining the ten parameters of the simulation model of the projection apparatus used. For this purpose, a linear equation system is set up, wherein for all Parameters of the simulation model of the projection apparatus the minimum Value of the standard deviation is determined. These parameters of the simulation model can be attributed to the projection apparatus to the respective Position error already correct during the exposure.

The residual error values in the x and y direction then become the Determination of a mean positional error used and provide the first offset values corrected for the systematic position errors the exposure fields of the first subset is.

Furthermore becomes a value from the residual error values in the x and y directions for the statistical fluctuation of the mean positional error by means of a Probability distribution determined. This is for example a normal distribution suitable, where for the statistical fluctuation the mean positional error the standard deviation of the residual error values be determined in each case in the x and y direction. This is not restrictive to understand, because the expert many probability distributions are known for This step can be used.

Up to this point of the procedure, the systematic positional errors were determined from the probability distribution using the simulation model and statistical fluctuation of the mean positional error. These sizes are used below for not using the overlay meter 22 measured second offset values.

The second offset value for each exposure field 5 of the second subset consists of second systematic position errors and the mean positional error together. The contribution of the second systematic bias will be based on the Simulation model for determines the exposure fields of the second subset. To do this the coordinate values of the exposure fields of the second subset in the calculation rules for the systematic positional error Ovl_X (x, y) or Ovl_Y (x, y) is used.

For this purpose, the values Ovl_X (x, y) and Ovl_Y (x, y) of the systematic position errors in the x and y directions are calculated at at least four positions within the exposure field. From these four values for Ovl_X (x, y) and Ovl_Y (x, y) will be hereafter
the maximum values of the systematic position errors in the x- and y-direction respectively. However, it is also intended, for example, to use the mean value or another suitable variable, eg weighted average, geometric mean value, smallest value or the like

The determination of the second offset value in the x direction is carried out according to the following equation: B_X = Ovl_X (x, y) + 3 · SIGMA.

there represent x the position of the exposure field, B_X the second offset value in the x direction, Ovl_X (x, y) that determined from the simulation model systematic position error in the x-direction and SIGMA the standard deviation SIGMA the residual error values. The multiplication of the standard deviation the residual error value with the factor of three has in practice achieved the best results. However, it should not be explicitly excluded Other factors to be used, for example, in the Range from one to five can. For probability distributions, the are not normally distributed, a similar size can be specified. It is within the scope of the invention, the statistical contribution, which is represented here by the standard deviation SIGMA, from arbitrary Distributions or to be determined by adaptation to measured values.

Analogously, the determination of the second offset value in the y direction is carried out using the following equation: B_Y = Ovl_Y (x, y) + 3 · SIGMA.

there represent y the position of the exposure field, B_Y the second offset value in the y direction, Ovl_Y (x, y) that determined from the simulation model systematic misalignment in y-direction and sigma the standard deviation the residual error values.

The result of the calculation of the second offset values in the x-direction and the y-direction is shown graphically in FIG 4 illustrated. The calculated second offset values are represented in this illustration by overlay targets whose first and second measurement marks have corresponding second offset values. For exposure fields that belong to the first subset, a corresponding overlay target is also drawn. It is also provided in the context of the invention that the second subset is expanded by the exposure fields of the first subset, so that all the exposure fields on the semiconductor wafer 10 are each represented by an offset value, which was obtained on the basis of the aforementioned calculation rules.

From the second offset values B_X and B_Y, the following is a distribution or histogram 40 the second offset values formed. The histogram 40 The first offset values can also be taken into account or extended by the offset values calculated for the exposure fields of the first subset analogously to the second offset values, so that the second offset values are determined for the entire semiconductor wafer. A typical histogram 40 is in 5 shown.

To a simple determination of the histogram 40 the first offset values and the second offset values, a data field is provided with first memory locations. The first memory locations are suitable for storing at least one numerical value. The content of each memory location includes the number of first offset values or second offset values assigned to it. For this purpose, a range of offset values is assigned to each first memory location. The data field thus covers offset values within a predetermined lower limit and a predetermined upper limit. By assigning each first offset value and second offset value to one of the first memory locations, the first offset values and second offset values are stored as in a histogram represented by the memory cells.

For example, in the fabrication of semiconductor memories, the range of offset values relevant in modern manufacturing processes includes the lower limit of about 5 nm and the upper limit of about 25 nm. The range of offset values assigned to a first memory location is about 0.25, for example nm. In 5 becomes the histogram 40 according to these lower limit and upper limit values shows.

in the Of course, it is also provided under the invention, the perform first memory cells so that the first offset values or the second offset values respectively separate for a proportion in the x-direction for stored a proportion in the y direction can be.

From the histogram 40 is a frequency distribution below 42 certainly. The frequency distribution 42 serves to determine what proportion 44 the exposure fields within a given offset value 46 lies. In order to determine this proportion, in one possible embodiment the data field is extended by second storage locations.

The second memory locations comprise for their area the sum of the number of first offset values or second offset values assigned from the lower limit. This will produce a cumulative histogram or frequency distribution 42 which is normalized in a subsequent step to the total number of first offset values and second offset values. The second memory cells are also suitable for storing the number of first offset values or second offset values assigned to them separately for a portion in the direction of a first coordinate axis and one for a portion in the direction of a second coordinate axis. The cumulative and normalized frequency distribution 42 is in 6 shown, where the frequency distribution 42 from the histogram 40 according to 5 was calculated.

Around to carry out the standardization, becomes, for example, a counter is provided, which is suitable, the total number of first offset values and second offset values.

These Total number then becomes the normalization of the contents of the second memory cells used.

In 7 is in an enlarged view the relevant area for the determination of the share 44 the exposure fields within the default offset value 46 shown. In this example, the default offset value is 46 about 20 nm, which is required by way of example that the predetermined proportion 44 the exposure fields within the predetermined offset value is at least 99%. as in 7 This semiconductor wafer would be recognizable 10 meet this overlay specification and can be added to the further process steps of manufacturing.

Those semiconductor wafers 10 whose first offset values and second offset values have a lower portion are subjected to post-processing. The post-processing includes, for example, the removal of the resist layer. Subsequently, another resist layer is applied to the front side of the semiconductor wafer 10 and then the circuit pattern is projected again into the exposure fields on the further resist layer.

In 8th are summarized the essential process steps in the determination of positional errors in the photolithographic projection, in particular in the structuring of a semiconductor wafer with a circuit pattern, which represent the first aspect of the invention.

In step 100 the provision of a semiconductor wafer takes place 10 with one on a front side of the semiconductor wafer 10 applied resist layer 14 ,

Subsequently, in step 102 providing at least two circuit levels for projecting each of a circuit pattern to form an integrated circuit.

After providing a projection apparatus 5 in step 104 and providing a simulation model for the projection apparatus containing a calculation rule 5 in step 106 Splits the front of the semiconductor wafer 10 in a multiple array of exposure fields 26 ,

Subsequently, in step 108 the circuit pattern successively in exposure fields 26 on the resist layer 14 projected.

Then in the step 110 splitting the exposure fields 26 in a first subset 36 and into a second subset.

Subsequently, in the step 112 first offset values for each overlay target 30 in each exposure field of the first subset, the first offset values being composed of first bias errors and a mean pitch error.

Thereafter, the systematic positional errors are determined in each exposure field 26 the first subset 36 on the basis of a simulation model of the projection apparatus 5 in step 114 and determining a mean positional error based on the first offset values of the exposure fields of the first subset corrected by the systematic position errors in the step 116 , Subsequently, in step 118 a second offset value for each exposure field 5 the second subset, which consists of second systematic errors and the middle Position error composed, determined, wherein the contribution of the second systematic position errors based on the simulation model for the exposure fields of the second subset is determined.

Finally, in the step 120 determining a frequency distribution 42 the first offset values and the second offset values to determine what fraction 44 the exposure fields within a given offset value 46 lies.

In 9 are summarized the main process steps in the evaluation of measurement results of an overlay measuring device, which represent the second aspect of the invention.

In step 130 providing an input data set comprising a plurality of first offset values, wherein the first offset values are from measurements of a first subset of overlay targets of two circuit levels arranged in a multiple array of exposure fields 26 on a semiconductor wafer 10 with a projection apparatus 5 were applied, determined and composed of first systematic errors and a mean misalignment.

In step 132 becomes a calculation model containing simulation model for the projection apparatus 5 provided.

Subsequently, in the step 134 determining the systematic attitude errors in each exposure field 26 the first subset 36 on the basis of the simulation model of the projection apparatus 5 and then in the step 136 determining a mean positional error based on the first offset values of the exposure fields of the first subset corrected for the systematic position errors.

Subsequently, in step 138 a second offset value for each exposure field 26 a second subset which is different from the first subset and which is composed of second systematic position errors and the mean positional error, the contribution of the second systematic position errors being determined on the basis of the simulation model for the exposure fields of the second subset.

Finally, in the step 140 determining a frequency distribution 42 the first offset values and the second offset values to determine what fraction 44 the exposure fields within a given offset value 46 lies.

The method according to 9 is particularly suitable, the overlay meter 22 according to 2 to control. The overlay meter 22 , as in 2 shown, adjacent to the substrate holder 23 which is suitable for the semiconductor wafer 10 to pick up the light source 24 for illuminating the on the substrate holder 23 stored semiconductor wafer 10 and the microscope 26 which is suitable from the semiconductor wafer 10 reflected light of the light source 24 and to generate therefrom an input data set, a processing means 27 , The processing agent 27 the input data set is supplied to, wherein the processing means 27 is suitable, the method according to 9 perform.

The processing agent 27 For example, it includes one or more processors connected to a computer via a connection, such as a network. On a computer-readable storage medium, such as a magnetic disk or an optical disk, the instructions are stored that allow the one or more processors, the method for controlling the overlay meter 22 perform.

5

projector

10

Semiconductor wafer

12

substrate holder

14

resist layer

16

light source

18

reticle

20

projection lens

22

Overlay meter

24

Further light source

26

microscope

27

processing means

28

exposure field

30

Overlay target

32

first measuring mark

34

second measuring mark

36

first subset

38

residual error value

40

histogram

42

frequency distribution

44

proportion of

46

maximum offset value

100-138

steps

Claims (46)

Method for determining positional errors in photolithographic projection, in particular in the structuring of a semiconductor wafer having a circuit pattern, comprising the following steps: - providing a semiconductor wafer ( 10 ) with a on a front side of the semiconductor wafer ( 10 ) applied resist layer ( 14 ); Providing at least two circuit levels for projecting a respective circuit pattern to form an integrated circuit, wherein for the first circuit level several first measuring marks ( 32 ), which are suitable, together with second measuring marks ( 34 ) of the second circuit level in each case an overlay target ( 30 ) to build; - Provision of a projection apparatus ( 5 ); Providing a simulation model for the projection apparatus containing a calculation specification ( 5 ); Dividing the front side of the semiconductor wafer ( 10 ) into a multiple arrangement of exposure fields ( 26 ); Successive projecting of the circuit pattern into exposure fields ( 26 ) on the resist layer ( 14 ) to form a resist pattern; - splitting the exposure fields ( 26 ) into a first subset ( 36 ) and a second subset; Determining first offset values for each overlay target ( 30 in each exposure field of the first subset, the first offset values being composed of first bias errors and a mean pitch error; Determine the systematic position errors in each exposure field ( 26 ) of the first subset ( 36 ) based on a simulation model of the projection apparatus ( 5 ); Determining a mean positional error based on the first offset values of the exposure fields of the first subset corrected by the systematic position errors; Determining a second offset value for each exposure field ( 5 ) the second subset, which is composed of second systematic position errors and the mean position error, the contribution of the second systematic position errors being determined on the basis of the simulation model for the exposure fields of the second subset; and determining a frequency distribution ( 42 ) of the first offset values and the second offset values to determine what fraction ( 44 ) of the exposure fields within a predetermined offset value ( 46 ) lies. The method of claim 1, wherein the step of providing the circuit pattern comprises the circuit pattern surrounded by a substantially rectangular saw frame and for the first circuit plane at least four first measuring marks are provided, which are suitable together with at least four second measuring marks of the second Circuit level one overlay target each ( 30 ), wherein the at least four overlay targets ( 40 ) are arranged at each corner of the saw frame. Method according to one of claims 1 or 2, wherein the step of providing the projection apparatus ( 5 ), that for the projection apparatus ( 5 ) a wafer scanner is provided. The method of claim 3, wherein the step of Providing the simulation model includes that of a simulation model provided with at least ten parameters that are appropriate is a translation error, a grid rotation error, a grid magnification error, a field rotation error and a field magnification error each in two mutually perpendicular coordinate axes specify. Method according to Claim 4, in which the following calculation is carried out to determine systematic positional errors with the simulation model: Ovl_X (x, y) = TransX + CF (x) * GMagX - CF (y) * GRotY + IF (x) * FMagX - IF (y) * FRotY, and Ovl_Y (x, y) = TransY + CF (y) * GMagY + CF (x) * GRotX + IF (y) * FMagY + IF (x) * FRotX; where x is the position of the exposure field along the first coordinate axis, y the position of the exposure field along the second coordinate axis, Ovl_X (x, y) or Ovl_Y (x, y) the systematic position error, TransX or TransY parameters of the translation error, GMagX or GMagY parameters of the grid magnification error, GRotX or GRotY parameters of the grid rotation error, FMagX or FMagY parameters of the field magnification error and FRotX or FRotY parameters of the field rotation error in the x-direction and y-direction respectively and CF or IF represent a decomposition of the coordinate values into a contribution on the semiconductor wafer or within an exposure field. Method according to one of claims 1 to 5, wherein the step of determining first offset values for each overlay target ( 30 ) in each exposure field of the first subset comprises: - providing an overlay measuring device ( 22 ); and measuring the overlay targets ( 30 ) in each exposure field of the first subset with the overlay measuring device ( 22 ). Method according to one of claims 1 to 6, wherein the first Offset values are determined in each case in the x-direction and y-direction. Method according to one of claims 1 to 7, wherein in addition for each Overlay target in each exposure field of the first subset the first offset values and the systematic position errors residual error value for each overlay target in each exposure field of the first subset is determined. Method according to Claim 8, in which the residual error value is calculated according to the following equations: Res_X = M_X - Ovl_X (x, y) and Res_Y = M_Y - Ovl_Y (x, y), where Res_X or Res_Y the residual error value in the x- or y-direction and M_X or M_Y the first offset value in x- or Represent y direction. Method according to one of claims 8 or 9, wherein in addition the standard deviation of the residual error value in the x and y direction is determined, which subsequently for determining the ten parameters of the simulation model of the projection apparatus be used by the minimum value of the standard deviation is determined. The method of claim 8 to 10, wherein the step of determining the mean positional error by the order of the systematic Positional error corrected first offset values of the exposure fields The first subset includes that the mean of the residual Error values in x- and y-direction determined as mean positional error becomes. Method according to one of claims 8 to 11, wherein in addition from the residual error values in the x and y direction, a value for the statistical fluctuation the mean positional error based on a probability distribution is determined. The method of claim 12, further comprising the probability distribution is a normal distribution and for the statistical fluctuation of the mean positional error the standard deviation the residual error values are determined in each case in the x and y direction become. Method according to one of claims 5 to 13, wherein the step determining the second offset values for each exposure field of second subset includes: - Calculate the values Ovl_X (x, y) or Ovl_Y (x, y) of the systematic position error in the x and y direction at least four positions within the exposure field; - Determine the maximum values of the systematic position errors in each case in and y-direction; - Determine of the second offset value in the x-direction using the following equation: B_X = Ovl_X (x, y) + 3 · SIGMA, where x is the position of the exposure field, B_X is the second offset value in the x direction, Ovl_X (x, y) that determined from the simulation model systematic positional error in the x-direction and SIGMA represent the standard deviation of the residual error values; and - Determine the second offset value in the y direction using the following equation: B_Y = Ovl_Y (x, y) + 3 · SIGMA, where y is the position of the exposure field, B_Y is the second offset value in the y direction, Ovl_Y (x, y) that determined from the simulation model systematic positional error in the y direction and SIGMA represent the standard deviation of the residual error values. Method according to one of claims 1 to 14, wherein the step splitting the exposure fields into the first subset and the second subset comprises the first subset about 10% of the Exposure fields includes. Method according to one of claims 1 to 15, wherein the step the provision of the circuit levels includes that a circuit design of a Semiconductor memory with dynamic memory cells comprising trench capacitors is provided, the smallest in the trench capacitor Has dimensions of 100 nm or less. The method of claim 16, wherein the step of determining the frequency distribution of the first offset values and the second offset values comprises the predetermined offset value (16). 46 ) is about 20 nm. The method of claim 17, wherein the step of determining the frequency distribution of the first offset values and the second offset values comprises that the predetermined fraction ( 44 ) of the exposure fields within the predetermined offset value is at least 99%. The method of claim 18, wherein for semiconductor wafers, their first offset values and second offset values are lower Share, the following step is carried out: - Feeding the Semiconductor wafer for a post-processing. The method of claim 19, wherein the post-processing step comprises: removing the resist layer; Applying a further resist layer to the front side of the semiconductor wafer ( 10 ); and re-projecting the circuit pattern into the exposure fields on the further resist layer. A method of evaluating measurement results of an overlay meter comprising the steps of: providing an input data set comprising a plurality of first offset values, wherein the first offset values are from measurements of a first subset of overlay targets of two circuit levels arranged in a multiple array of exposure fields ( 26 ) on a semiconductor wafer ( 10 ) with a projection apparatus ( 5 ), were determined and are composed of first systematic errors and a mean misalignment; - Provide a computational regulation ent holding simulation model for the projection apparatus ( 5 ); Determine the systematic position errors in each exposure field ( 26 ) of the first subset ( 36 ) on the basis of the simulation model of the projection apparatus ( 5 ); Determining a mean positional error based on the first offset values of the exposure fields of the first subset corrected by the systematic position errors; Determining a second offset value for each exposure field ( 26 ) a second subset, which is different from the first subset, and which is composed of second systematic position errors and the mean positional error, wherein the contribution of the second systematic position errors is determined on the basis of the simulation model for the exposure fields of the second subset; and determining a frequency distribution ( 42 ) of the first offset values and the second offset values to determine what fraction ( 44 ) of the exposure fields within a predetermined offset value ( 46 ) lies. The method of claim 21, wherein the step Providing the simulation model includes that a simulation model provided with at least ten parameters that are appropriate is a translation error, a grid rotation error, a grid magnification error, a field rotation error and a field magnification error each in two mutually perpendicular coordinate axes specify. Method according to Claim 22, in which the following calculation is carried out for determining systematic positional errors on the basis of the simulation model: Ovl_X (x, y) = TransX + CF (x) * GMagX - CF (y) * GRotY + IF (x) * FMagX - IF (y) * FRotY, and Ovl_Y (x, y) = TransY + CF (y) * GMagY + CF (x) * GRotX + IF (y) * FMagY + IF (x) * FRotX; where x is the position of the exposure field along the first coordinate axis, y the position of the exposure field along the second coordinate axis, Ovl_X (x, y) or Ovl_Y (x, y) the systematic position error, TransX or TransY parameters of the translation error, GMagX or GMagY parameters of the grid magnification error, GRotX or GRotY parameters of the grid rotation error, FMagX or FMagY parameters of the field magnification error and FRotX or FRotY parameters of the field rotation error in the x-direction and y-direction respectively and CF or IF represent a decomposition of the coordinate values into a contribution on the semiconductor wafer or within an exposure field. The method of any of claims 21 to 23, wherein the step of determining first offset values for each overlay target ( 30 ) in each exposure field of the first subset comprises: - providing an overlay measuring device ( 22 ); and measuring the overlay targets ( 30 ) in each exposure field of the first subset with the overlay measuring device ( 22 ). A method according to any one of claims 21 to 24, wherein the first offset values in each case in the x-direction and y-direction determined become. A method according to any one of claims 21 to 25, further comprising for each Overlay target in each exposure field of the first subset of the first offset values and the systematic positional errors a residual Error value for each overlay target in each exposure field of the first subset is determined. The method of claim 26, wherein the residual Error value according to the following Equations is calculated: Res_X = M_X - Ovl_X (x, y) and Res_Y = M_Y - Ovl_Y (x, y), where Res_X or Res_Y the residual error value in the x and y direction and M_X or M_Y represent the first offset value in the x or y direction. A method according to any one of claims 26 or 27 further comprising the standard deviation of the residual error value in the x and y direction is determined, which subsequently for determining the ten parameters of the simulation model of the projection apparatus be used by the minimum value of the standard deviation is determined. The method of claim 26 to 28, wherein the step of determining the mean positional error by the order of the systematic Positional error corrected first offset values of the exposure fields The first subset includes that the mean of the residual Error values in x- and y-direction determined as mean positional error becomes. A method according to any one of claims 26 to 29, further comprising from the residual error values in the x and y direction, a value for the statistical fluctuation the mean positional error based on a probability distribution is determined. The method of claim 30, further comprising the probability distribution is a normal distribution and for the statistical fluctuation of the mean positional error the standard deviation the residual error values are determined in each case in the x and y direction become. Method according to one of claims 23 to 31, in which the step of determining the second offset values for each exposure field of the second subset comprises: - calculating the values Ovl_X (x, y) and Ovl_Y (x, y) of the systematic position errors in the x and y directions to at least four Positions within the exposure field; Determining the maximum values of the systematic position errors in the x and y direction in each case; Determining the second offset value in the x direction using the following equation: B_X = Ovl_X (x, y) + 3 * SIGMA, where x is the position of the exposure field, B_X is the second offset value in the x direction, Ovl_X (x, y) is the the systematic error in the x-direction determined by the simulation model and SIGMA represent the standard deviation of the residual error values; and determining the second offset value in the y direction using the following equation: B_Y = Ovl_Y (x, y) + 3 * SIGMA, where y is the position of the exposure field, B_Y is the second offset value in the y direction, and Ovl_Y (x, y) is the from the simulation model, systematic positional errors in the y direction and SIGMA represent the standard deviation of the residual error values. A method according to any one of claims 21 to 32, wherein the Step of dividing the exposure fields into the first subset and the second subset comprises the first subset comprising about 10% of the Exposure fields includes. The method of claim 33, wherein the step of determining the frequency distribution of the first offset values and the second offset values comprises the predetermined offset value (16). 46 ) is about 20 nm. The method of claim 34, wherein the step of determining the frequency distribution of the first offset values and the second offset values comprises that the predetermined fraction ( 44 ) of the exposure fields within the predetermined offset value is at least 99%. A method according to any one of claims 21 to 35, wherein the Step of determining the frequency distribution the first offset values and the second offset values comprises: - Provide a data field with first memory locations that are suitable storing at least one numerical value, each first memory location in each case a range of offset values is assigned, so that the Data field offset values within a given lower limit and a predetermined upper limit; and - To assign each first offset value and second offset value to a memory location, the contents of the memory location being the number of their assigned ones comprises first offset values or second offset values. The method of claim 36, wherein the lower limit an offset value of approximately 5 nm and the upper limit an offset value of about 25 nm represent and wherein the area of a memory location represents about 0.25 nm. A method according to any one of claims 36 to 37, wherein the first memory cells above are the number of first offset values assigned to it or second offset values each separately for a fraction in the direction a first coordinate axis and one for a proportion in the direction of a store second coordinate axis. A method according to any one of claims 36 to 38, wherein the Data field above also comprises second memory cells, each of a first memory location correspond, wherein the content of the second memory location is the sum the contents of those first memory cells comprises, the same or assigned to smaller first offset values or second offset values were. The method of claim 39, wherein the second Memory cells above are the number of first offset values assigned to it or second offset values each separately for a fraction in the direction a first coordinate axis and one for a portion in the direction store a second coordinate axis. The method of claim 40, wherein a counter is provided which is suitable to the total number of first offset values and determine second offset values, and wherein the content of the second Memory cells is normalized to the total number. Overlay measurement device which includes: - one Substrate holder adapted to receive a semiconductor wafer; - a light source for illuminating a semiconductor wafer deposited on the substrate holder; and - one Microscope, which is suitable, reflected from the semiconductor wafer Light of the light source to detect and from an input data set to be supplied to a processing means, wherein the processing means is suitable to carry out a method according to any one of claims 21 to 35. Overlay meter according to claim 42, wherein the processing means comprises one or more processors includes. A computer readable storage medium having stored thereon instructions that enable one or more processors to execute a method according to any of claims 21 to 41 ren. A storage medium according to claim 44, which is a magnetic disk includes. A storage medium according to claim 44, comprising an optical disk includes.