IMPROVED METHOD FOR MEASURING THE ALIGNMENT OF A SMALL LINE DEFECTION ERROR BY T-SAXS TRANSMISSION; CORRESPONDING INSTRUMENTAL SYSTEM AND COMPUTER PROGRAM PRODUCT

Description

[0001] The present invention relates to a method for measuring an error in the superposition of two line networks during the manufacture of a microelectronic component.

[0002] The present invention relates more particularly to a measuring method implementing a technique of small angle X-ray scattering by transmission, or T-SAXS (Transmission Small Angle X-ray Scattering).

[0003] The manufacture of a microelectronic component requires that the different levels constituting it be precisely superimposed, so that the patterns carried by these different levels are correctly positioned in relation to each other in order to guarantee the correct functioning of the final component.

[0004] It is known to control the precision of the superposition of two levels by measuring an alignment error, or "overlay" (abusively using the English term as a person skilled in the art would do), between a first network of lines provided on the first level and a second network of lines provided on the second level.

[0005] To measure this overlay, document US 9,885,962 B2 presents a method using a T-SAXS technique. This involves illuminating the first and second arrays of lines superimposed on each other with an X-ray beam and evaluating the overlay from the diffraction patterns obtained by transmission for different angles of incidence of the X-ray beam.

[0006] More precisely, for a given angle of incidence, the diffraction pattern of the superposition of the first and second gratings of lines presents a succession of maxima, or Bragg peak. Each Bragg peak is identified by an integer n, called the Bragg order.

[0007] The overlay measurement method consists, by successive measurements, of reconstructing a curve giving the intensity of the Bragg peak of order n for different values ​​of the angle of incidence, as well as a curve giving the intensity of the Bragg peak of order n' for different values ​​of the angle of incidence.

[0008] The overlay measurement process continues by determining the position of an intensity minimum on the curve associated with Bragg order n, as well as the position of the corresponding intensity minimum on the curve associated with Bragg order n'.

[0009] Finally, the method for measuring the overlay provides for the calculation of the ratio of the difference between the positions of the corresponding minima between the Bragg orders n and n' and the difference between the position of the Bragg peaks n and n'. According to document US 9,885,962 B2, this difference is none other than the desired overlay.

[0010] However, this last statement is not correct since it is also necessary to know the spacing between the first and second line networks in order to determine the overlay from the ratio. However, this spacing, which can differ significantly between its theoretical value and its actual value, influences the final measured value of the overlay. It is therefore necessary that the overlay measurement method also provides for the measurement of the actual spacing between the first and second line networks.

[0011] In addition to this problem, this state-of-the-art method remains difficult to implement on a microelectronic component manufacturing line, because it requires not only the measurement of the intensity for different angles of incidence, but above all the repetition of these measurements for several different Bragg peaks, in order to reconstruct as many intensity curves and then be able to compare them two by two.

[0012] While this method can possibly be implemented when the component being manufactured only has a small number of overlay evaluation zones (areas carrying the line networks in order to measure the overlay), it becomes particularly difficult to implement when the component being manufactured has a large number of overlay evaluation zones. This is becoming the norm, for example for the manufacture of silicon wafers, for which it is necessary to ensure perfect superposition of the patterns at all points of the wafer, which is controlled by multiplying the overlay measurement zones.

[0013] The method presented in document US 9,885,962 B2 is therefore not practical.

[0014] The aim of the present invention is to propose a particularly simple method for measuring the overlay by T-SAXS.

[0015] For this purpose, the subject of the invention is a method for measuring an alignment deviation between a first line array carried by a first level of a microelectronic component and a second line array carried by a second level of the microelectronic component, an orthonormal xyz reference frame being associated with the microelectronic component, the lines of the first and second line arrays being oriented in the y direction, the first and second line arrays being superimposed on each other in the z direction, and the alignment deviation being measured in the x direction, the measurement method implementing a T-SAXS transmission X-ray scattering technique, characterized in that the method comprises the steps of: acquiring from the microelectronic component a plurality of intensity measurements of a transmitted X-ray beam for a plurality of angles of incidence of the X-ray beam, the angle of incidence being defined as the angle,evaluated in the plane defined by the x and z directions, between the z direction and the direction of incidence of the X-ray beam; reconstruct, from the plurality of intensity measurements, a diffraction pattern according to spatial frequencies, Q x And Q z , associated respectively with the x and z directions; determine, on the diffraction pattern, a position according to the spatial frequency Q x of a Bragg rod of interest and the positions according to the spatial frequency Q z of each point of a pair of characteristic points of the Bragg rod of interest; and, calculate the alignment deviation from the determined positions.

[0016] According to particular embodiments, the method comprises one or more of the following characteristics, taken in isolation or in all technically possible combinations: the pair of characteristic points of the Bragg rod of interest consists of the minimum of index i and the corresponding minimum of index -i of the Bragg rod of interest, i being an integer greater than or equal to unity. the OVL alignment deviation is calculated using the relation: OVL = 1 + 2 i − 1 Q z n , i + Q z n , − i Q x n . Q z n , i − Q z n , − i , where n is the Bragg order of the Bragg rod of interest, i is the index of the pair of minima considered, Q x n is the position according to the spatial frequency Q x of the Bragg rod of interest, Q z n , i is the position according to the spatial frequency Q z of the minimum of index i and Q z n , − i is the position according to the spatial frequency Q z of the corresponding minimum of index -i. determine a position according to the spatial frequency Q x of the Bragg rod of interest involves the determination of positions according to the spatial frequency Q z of each point of another pair of characteristic points of the Bragg rod of interest from the diffraction pattern. said other pair of characteristic points consists of the minimum of index j and the corresponding minimum of index j of the Bragg rod of interest, the index j being different from the index i. the pair of characteristic points of the Bragg rod of interest consists of the maximum of index i and the corresponding maximum of index -i of the Bragg rod of interest. the step of determining, on the diffraction pattern, a position according to the spatial frequency Q x of a Bragg rod of interest and the positions according to the spatial frequency Q z of each point of a pair of characteristic points of the Bragg rod of interest is iterated for another pair of characteristic points of the Bragg rod of interest and / or for another Bragg rod of interest, the alignment deviation resulting from the average of the alignment deviations calculated for each iteration.

[0017] The invention also relates to an instrumental system of the T-SAXS type, comprising a detector, acquisition electronics and a computer, characterized in that the computer is suitably programmed so that said instrumental system implements a method for measuring an alignment deviation according to the preceding method.

[0018] Preferably, the system comprises a detector adapted to measure the intensity of the transmitted beam over a reduced diffraction angle range making it possible to track the Bragg rod of interest during the acquisition step.

[0019] The invention also relates to a computer program product comprising software instructions which, when executed by a computer of the preceding instrumental system, allow the latter to implement a method in accordance with the preceding method.

[0020] The invention and its advantages will be better understood upon reading the following detailed description of a particular embodiment, given solely as a non-limiting example, this description being made with reference to the appended drawings in which: [ Fig 1 ] There figure 1 is a schematic representation of an instrumental system for measuring overlay by T-SAXS; [ Fig 2 ] There figure 2 is a representation of the Fourier transform of a pattern for which the overlay is zero; [ Fig 3 ] There figure 3 is a representation of the Fourier transform of a pattern for which the overlay is non-zero; [ Fig 4 ] There figure 4 is a block representation of an embodiment of the overlay measurement method according to the invention; [ Fig 5 ] There figure 5 is a graph representing the intensity as a function of the diffraction angle for different values ​​of the angle of incidence for a sample with a single line network, i.e. without overlay but allowing the technique to be illustrated; [ Fig 6 ] There figure 6 is the Fourier transform reconstructed from the measurements of the figure 5 ; And, [ Fig 7 ] There figure 7 is an enlargement of the figure 3 in the vicinity of the first-order Bragg peak on which position measurements are made from which to determine the overlay.

[0021] There figure 1 represents an instrumental system 1 for measuring the overlay.

[0022] An X-ray source S emits an X-ray beam in a direction z 0 , perpendicular to an observation plane P. The incident beam falls at an origin point O of the observation plane P. Directions x 0 and y 0 define an orthonormal reference frame of the plane P.

[0023] A microelectronic component C is interposed between the source S and the observation plane P.

[0024] Component C has a Z zone for determining the overlay.

[0025] The center A of the surface of zone Z is placed on the axis of incidence of the X-ray beam.

[0026] An orthonormal xyz coordinate system is associated with component C. This coordinate system is attached to the center A of the surface of area Z so that the y direction corresponds to the y 0 direction, and the z direction is normal to the surface of area Z.

[0027] As specified below, the lines of the line array(s) used to measure the overlay are arranged parallel to the y 0 axis, so as to have Bragg peaks along the x 0 axis.

[0028] The component C is placed on a support 30 allowing the component C to be rotated around the direction y so as to modify the angle ϕ of incidence of the X-ray beam on the zone Z, that is to say the angle between the direction z and the direction z 0 , which is also the angle between the direction x and the direction x 0 .

[0029] A detector 10 is placed in the observation plane P. It is for example composed of a bar of sensors arranged in the direction x0.

[0030] The intensity at point B of the x-axis 0 , measured by the sensor located at point B, depends on the angle 2 θ between the AO direction and the AB direction. This intensity is noted I ( θ ).

[0031] As is known, the image in the observation plane P is related to the spatial Fourier transform of the area Z illuminated by the incident beam.

[0032] In reciprocal space, the coordinate associated with the x 0 direction is the spatial frequency Q 0 defined by: Q 0 = 2 sin θ λ

[0033] That is, still in the reciprocal space, but considering the spatial frequencies Qx and Qz respectively associated with the directions x and z of the frame linked to the component C (and consequently to the zone Z): Q x = Q 0 cos φ Q z = Q 0 sin φ

[0034] The detector 10 is connected to an electronic device, shown schematically in the figure 1 by a cube bearing the reference 20.

[0035] The device 20 comprises control electronics making it possible to control the support 30 so that it positions the component C according to a set value of the angle of incidence.

[0036] The device 20 also includes acquisition electronics making it possible to carry out suitable pre-processing on the signals delivered by each of the sensors of the detector 10 and to digitize them.

[0037] The device 10 further comprises a calculator for processing the pre-processed and digitized signals. The calculator is a computer comprising calculation means, such as a processor, and storage means, such as a memory. The memory stores in particular the instructions of computer programs, in particular a program whose execution allows the implementation of the overlay measurement method according to the invention.

[0038] From a theoretical point of view, it is possible to calculate exactly the spatial Fourier transform of a pattern.

[0039] So the left part of the figure 2 represents, in the direct space of the x and z coordinates, a fraction of the pattern carried by the zone Z of the component C. This is the superposition, in the thickness of the circuit (i.e. in the z direction), of a first network of lines 40 and a second network of lines 50. For example, two lines 41 and 42 of the first network and two lines 51 and 52 of the second network are represented. The lines of each of these networks extend in the y direction.

[0040] A line has a roughly rectangular section in the xy plane, of width l and depth p. Two lines of a network are spaced apart by a pitch d.

[0041] The two networks of lines 40 and 50 are separated from each other by a distance D.

[0042] On the figure 2 , the first and second line networks are perfectly superimposed along the x direction.

[0043] The exact calculation of the spatial Fourier transform of this pattern leads to the figure shown on the right part of the figure 2 . It is therefore a representation in the reciprocal space of the spatial frequencies Qx and Qz, respectively conjugated to the coordinates x and z.

[0044] For a zero value of Q z , the Fourier transform presents, according to the direction Q x , a succession of maxima, or Bragg peak. Each Bragg peak is identified by an integer n, called the Bragg order. The Bragg peak of order n is positioned at Q x n .

[0045] For a given Bragg peak, that is, a Q x given, the intensity according to the direction Q z constitutes what is called the nth-order Bragg rod. Whatever n, the nth-order Bragg rod exhibits alternating local maxima and minima. The nth-order Bragg rod is symmetrical about the axis Q x (has Q z = 0). Thus, the i th < minimum of the Bragg rod of order n, located at Q z n , i ( Q z n , i positive), corresponds to an i th< minimum of the Bragg rod of order n located at Q z n , − i = − Q z n , i .

[0046] We note that, when there is no overlay, the position along the axis Q z of the i th< minimum of the Bragg rod is constant, whatever the Bragg order n considered ( Q z n , i = Q z i ).

[0047] There figure 3 is similar to the figure 2 except that the first and second line arrays 40 and 50 are now slightly offset from each other along the x-direction of an OVL overlay. The lines of the second array 50 are angularly offset from the lines of the first array 40 by an angle a such as : tg α = OVL D

[0048] Or, as a first approximation, for values ​​of α small: α = OVL D

[0049] The exact calculation of the spatial Fourier transform of this pattern with overlay leads to the figure shown on the right part of the figure 3 . This is a representation in the reciprocal space of spatial frequencies Q x And Q z , respectively conjugates of the x and z coordinates.

[0050] We now see that the Bragg rods are essentially translated along the direction Q z of a quantity proportional to Qx, to within a constant.

[0051] This is particularly noticeable when following the first positive minimum of each Bragg rod as the Bragg order increases or when following the first negative minimum of each Bragg rod as the Bragg order increases.

[0052] We can thus write, for the first positive minimum of the Bragg rod of order n: Q z n , 1 = αQ x n + b

[0053] We can write, for the first negative minimum of the Bragg rod of order n: Q z n , − 1 = αQ x n − b

[0054] Therefore, for the same Bragg rod of order n, the sum of the positions of the first minima allows us to obtain α: Q z n , 1 + Q z n , − 1 = 2 αQ x n

[0055] Either : α = Q z n , 1 + Q z n , − 1 / 2 Q x n

[0056] Moreover, knowing that the constant b (position value according to direction Q z of the first positive minimum of the central Bragg rod, n=0) is none other than the inverse of the spacing parameter D (this results from the calculation by Fourier transformation), the difference in the positions of the first minima makes it possible to determine b .

[0057] In fact, we have: Q z n , 1 − Q z n , − 1 = 2 b

[0058] And : b = 1 / D = Q z n , 1 − Q z n , − 1 / 2

[0059] The overlay can then be determined from the values ​​of α and b: OVL = D . α = α b = Q z n , 1 + Q z n , − 1 Q x n . Q z n , 1 − Q z n , − 1

[0060] More generally, if we consider the pair of minima consisting of the i th < positive minimum and the i th < negative minimum, the previous equations become: Q z n , i = αQ x n + b . 1 + 2 i − 1 Q z n , − i = αQ x n − b . 1 + 2 i − 1

[0061] Either α : α = Q z n , i + Q z n , − i / 2 Q x n and b: b = 1 / D = Q z n , i − Q z n , − i / 2 1 + 2 i − 1

[0062] And therefore the overlay: OVL = D . α = α b = 1 + 2 i − 1 Q z n , i + Q z n , − i Q x n . Q z n , i − Q z n , − i

[0063] Even more generally, we could write equivalent relations for a pair of minima which would consist of the i th < minimum (positive or negative) and the j th < minimum (positive or negative).

[0064] Equivalent relations can also be determined for other characteristic points of the Bragg rod of order n, notably the intensity maxima of this rod.

[0065] Taking into account these theoretical considerations, an embodiment of the overlay measurement method according to the invention will now be presented with reference to the figure 4 .

[0066] The method 100 first consists of reconstructing at least the Bragg rod of order n considered as the rod of interest.

[0067] For this, in a first acquisition step 110, by means of the detector 10, the measurement of the intensity is iterated for different angles of incidence ϕ. For example, the angle of incidence is varied in steps of 1° in the range from +60° to -60° and for each value of the angle of incidence the intensity is measured. I ( θ ) as a function of the position along the x 0 axis of the observation plane P, i.e. as a function of the diffraction angle θ . We thus obtain a plurality of intensity curves, such as those represented on the figure 5 for three different values ​​of the angle of incidence.

[0068] The method 100 continues with a second step 120 consisting of reconstructing the diffraction pattern in the reciprocal space described by the coordinates Q x And Q z . The plurality of intensity curves obtained in step 110 and the relations (1) (2) and (3) above are used to obtain, for example, the figure of the figure 6 .

[0069] Once the Bragg rod of interest has been reconstructed, the method 100 continues with a step 130 of determining the position along the axis Q x of the Bragg rod of interest and the position along the axis Q z of a pair of characteristic points of the Bragg rod of interest.

[0070] For example, on the figure 7 , which is an enlarged representation of the figure 3 around the first-order Bragg rod as the Bragg rod of interest, we measure the position Q x 1 and positions Q z 1,1 of the first positive minimum and Q z 1 , − 1 of the first negative minimum along this Bragg rod. We obtain according to the example of the figure 7 : Q x 1 = 0.09 Q z 1,1 = 0.1033 Q z 1 , − 1 = − 0.0774

[0071] Alternatively, one could measure the position of the i th< positive and negative minima of the Bragg rod of interest. For example, for i=4 on the first Bragg rod, one obtains on the figure 7 : Q z 1,4 = 0.640857 Q z 1 , − 4 = − 0.61575

[0072] Finally, in a fourth step 140, the measurements of the positions of the minima are injected into relation (6) (or relation (7) for a minimum of index i greater than unity) to calculate the value of the overlay OVL of the zone Z of the component C analyzed.

[0073] For example, on the pair of first minimums, the value of overlay in the example of the figure 7 is 5nm.

[0074] The instrumental system 1 is adapted to implement the measurement method 100 just presented. In particular, the computer of the system 1 is suitably programmed to carry out the calculation steps, that is to say the steps other than the acquisition step.

[0075] Other embodiments of the process are possible.

[0076] Thus, alternatively, to avoid having to measure the position of the Bragg rod of interest, Q x n , one could combine the measurements of the positions of a first pair of minima (for example of order 1) and those of a second pair of minima (for example of order 4) of the Bragg rod of interest.

[0077] In another variant, other characteristic points of a Bragg rod can be considered: pair composed of minima of different index (whether these indices are positive and / or negative), pair composed of maxima of identical or different indices (whether these indices are positive and / or negative), pair composed of a maximum and a minimum of identical or different indices (whether these indices are positive and / or negative), etc. That is to say, in general, any correlation of points of the Bragg rod resulting from the translation of the latter due to the presence of an overlay.

[0078] To increase the accuracy of the overlay measurement, the process can be repeated for other pairs of characteristic points of the Bragg rod of interest (of order n) and / or another Bragg rod of interest (of order n' different from n). The different overlay measurements are then averaged to obtain a more accurate final value.

[0079] It should be noted that the lower the index n of the Bragg peak of interest, the greater the contrast, which improves the determination of the position of a local minimum of the intensity.

[0080] On the other hand, the larger the index n of the Bragg peak of interest, the greater the displacement of the minimum due to the overlay.

[0081] It is therefore necessary to make a compromise: if the criterion is the measurement time then it is better to work on a Bragg peak of low index n; on the other hand, if the criterion is the measurement resolution, then it is necessary to work on a Bragg peak of high index n.

[0082] Thus, in practice, the Bragg rod(s) of interest are selected with an order between 1 and 5, preferably equal to 3.

[0083] Moreover, the closer the pair of characteristic points used for the calculation of the overlay is to the value Q z = 0, the smaller the range of variation of the incidence angle can be. For example, for the pair of first minima of the Bragg rod of interest, the incidence angle can vary over a small range, between +10° and -10° for example. This reduces the time required for the first step of the process associated with collecting intensity measurements for different incidence angles.

[0084] Advantageously, since it is at least a question of following a single Bragg rod, for example that of the Bragg peak of order n (that is to say of considering a position Q x n fixed), the intensity can be measured over an angular range in 2 θ reduced corresponding to the values ​​of the diffraction angle θ given by, when ϕ varies over the incidence range: θ = asin λQ x n / 2 cos φ

[0085] If the incidence range is reduced, the angular range in 2 θ can also be reduced. Advantageously, the angular range in 2 θ is reduced to match the aperture of a single sensor.

[0086] In this case, the instrumental system can be simplified by using a detector comprising only a small number of sensors, or even a single sensor, suitably positioned in the observation plane P to be able to measure the intensity of the same Bragg peak of interest over the entire range of variation of the angle of incidence. For example, on the figure 5 , considering the Bragg peak of order 2 as the peak of interest, it would be sufficient to have in the observation plane P a single sensor whose aperture E would cover the angles 2 θ between 0.115 and 0.135°.

[0087] In the embodiment of the figure 1, it is the orientation of the component in the X-ray beam that is modified to vary the angle of incidence. Alternatively, the component is fixed and it is the source that is moved to vary the angle of incidence. The displacement of the source can be real or virtual using a system of adapted lenses.

Claims

1. A method (100) for measuring a misalignment (OVL) between a first array of lines (40) carried by a first level of a microelectronic component (C) and a second array of lines (50) carried by a second level of the microelectronic component, an orthonormal coordinate system xyz being associated with the microelectronic component, the lines of the first and second arrays of lines being oriented in the y direction, the first and second arrays of lines being superimposed one on the other in the z direction, and the misalignment (OVL) being measured in the x direction, the measurement method implementing a technique of transmission small angle X-ray scattering (T-SAXS), the method comprising the steps of: - acquiring, from the microelectronic component, a plurality of intensity measurements of a transmitted X-ray beam for a plurality of incidence angles of the X-ray beam, the incidence angle being defined as the angle, evaluated in the plane defined by the x and z directions, between the z direction and the direction of incidence of the X-ray beam; - reconstructing, from the plurality of intensity measurements, a diffraction pattern according to spatial frequencies Qx and Qz, associated respectively with the x and z directions; characterized in that the process comprises the steps of: - determining, on the diffraction pattern, a position according to the spatial frequency Qx of a Bragg rod of interest and the positions according to the spatial frequency Qz of each point of a pair of characteristic points of the Bragg rod of interest; and - calculating the misalignment from the determined positions.

2. The measuring method according to claim 1, wherein the pair of characteristic points of the Bragg rod of interest consists of the minimum of index i and the corresponding minimum of index -i of the Bragg rod of interest, i being an integer greater than or equal to unity.

3. The measuring method according to claim 2, wherein the misalignment OVL is calculated using the relationship: OVL = 1 + 2 i − 1 Q z n , i + Q z n , − i Q x n . Q z n , i − Q z n , − i where n is the Bragg order of the Bragg rod of interest, i is the index of the pair of minima taken into account, Q x n is the position according to the spatial frequency Qx of the Bragg rod of interest, Q z n , i is the position according to the spatial frequency Qz of the minimum of index i and Q z n , − i is the position according to the spatial frequency Qz of the corresponding minimum of index -i.

4. The measuring method according to claim 3, wherein determining a position according to the spatial frequency Qx of the Bragg rod of interest comprises determining the positions according to the spatial frequency Qz of each point of another pair of characteristic points of the Bragg rod of interest from the diffraction pattern.

5. The measuring method according to claim 4, wherein said other pair of characteristic points consists of the minimum of index j and the corresponding minimum of index -j of the Bragg rod of interest, the index j being different from the index i.

6. The measuring method according to claim 1, wherein the pair of characteristic points of the Bragg rod of interest consists of the maximum of index i and the corresponding maximum of index -i of the Bragg rod of interest.

7. The measuring method according to any one of the preceding claims, wherein the step of determining, on the diffraction pattern, a position according to the spatial frequency Qx of a Bragg rod of interest and the positions according to the spatial frequency Qz of each point of a pair of characteristic points of the Bragg rod of interest is iterated for another pair of characteristic points of the Bragg rod of interest and / or for another Bragg rod of interest, the misalignment resulting from the average of the misalignments calculated for each iteration.

8. An instrumental system (1) of the transmission small-angle X-ray scattering (T-SAXS) type, comprising a detector, acquisition electronics and a computer, characterized in that the computer is suitably programmed so that said instrumental system implements a method for measuring a misalignment according to any one of claims 1 to 7.

9. The instrumental system according to claim 8, comprising a detector adapted to measure the intensity of the beam transmitted over a reduced diffraction angle range making it possible to follow the Bragg rod of interest during the acquisition step.

10. A computer program product comprising software instructions which, when executed by a computer of an instrumental system conforming to any one of claims 8 and 9, allows the latter to implement a method according to any one of claims 1 to 7.