DE102014205660A1 - specimen - Google Patents

Abstract

The invention relates to a test body (10) for calibrating a Raman microscope with a substrate (12) of a first material and a calibration structure applied thereon, (14) of a second material different from the first material, wherein at least one of the first material and the second material is Raman active. According to the invention, it is provided that the calibration structure (14) has a thickness d1 of less than 30 nm

Description

The invention relates to a test specimen for calibrating a Raman microscope and to a method for producing a test specimen for calibrating a Raman microscope.

A Raman microscope is a microscope that uses the characteristic vibrational vibrations in molecules to image by detecting after excitation of the molecules by means of a laser, the photons emitted during relaxation and thus the Stokes lines are determined. Using a Raman microscope, the Raman spectra of the sample are recorded in addition to the light-optical image. It is worked with a raster method, so that the Raman spectra are spatially resolved. In this case, the examined area with the selected step size is moved line by line point by point under the microscope objective. This procedure is called "mapping". The Raman microscope thus combines optical microscopy and Raman spectroscopy.

Raman microscopes are currently used in particular to study biological systems at the cellular level, possibly also with time resolution. This way images of living cells can be generated. It is also possible via marker method z. B. to understand transport processes, which is relevant for clinical research, for example. Raman microscopy is thus a versatile measurement method for nondestructive characterization of z. B. chemical sample composition, crystallinity and mechanical stress influences. Since spatially resolved information in the micrometer range is obtained by stepwise scanning of the sample under the microscope, the quality of the images obtained is significantly dependent on the scanning device.

Therefore, a Raman microscope must be calibrated like any high-resolution scanning microscope. In particular, the sample stage must be calibrated to ensure that the desired position of the sample to be tested matches its actual position. A common method for calibrating microscopes is the use of a so-called standard or standard. Such a standard or normal are test specimens with precisely defined, measurable properties that are traced back to the international system of units (SI) through a completely closed calibration chain. The test specimen is measured with the microscope and the measurement results obtained are compared with the data known from the specimen and a calibration is performed.

To calibrate scanning devices on Raman microscopes, test specimens have generally been used which can also be used for other microscope types. Known test specimens, z. B. for atomic force microscopes, scanning electron microscopes or fluorescence microscopes u. a. produced by the company Piano and the company NT-MDT. However, the available 1D graticules with 120 nm deep indentations do not show sufficient Raman contrast between the upper and lower plateaus and additionally lead to disturbing edge effects.

The object of the invention is therefore to provide a test specimen which allows an increased accuracy in the calibration of a Raman microscope, wherein the calibration can be carried out without additional equipment.

The invention achieves the object by a test body having a substrate made of a first material, and a calibration structure applied thereon of a second material different from the first material, wherein at least one of the first material and the second material is Raman active and wherein the calibration structure is a Thickness d ₁ of less than 30 nm.

• a. Aufbringen einer Schicht eines photosensitiven oder elektronenstrahlsensitiven Lacks auf einem Substrat, das aus einem ersten Material besteht,
• b. Belichten der Lackschicht mit einem zur Kalibrierung eines Raman-Mikroskops geeigneten Muster,
• c. Aufbringen einer Schicht einer Dicke d₁ von unter 30 nm eines von dem ersten Material verschiedenen Materials auf das Substrat, wobei das erste Material und/oder das zweite Material Raman-aktiv ist,
• d. Entfernen der Schicht des photosensitiven Lacks.

The object is further achieved by a method for producing a Raman test specimen comprising the steps:

• a. Applying a layer of a photosensitive or electron beam-sensitive lacquer to a substrate which consists of a first material,
• b. Exposing the lacquer layer with a pattern suitable for the calibration of a Raman microscope,
• c. Depositing a layer of thickness d ₁ of less than 30 nm on the substrate of a material other than the first material, wherein the first material and / or the second material is Raman-active,
• d. Removing the layer of the photosensitive varnish.

Furthermore, the object is achieved by the use of a test body according to the invention for calibrating a Raman microscope.

The invention is based on the finding that a calibration of higher accuracy can be achieved by utilizing the Raman spectrometer integrated in the Raman microscope. According to the invention, a test specimen consists of a substrate made of a first material and a calibration structure applied thereon of a second material different from the first material, wherein at least one of the first material and the second material is Raman active and wherein the calibration structure has a thickness d ₁ of less than 30 nm. The specimen is thus not structured as known specimens with respect to its topography or fluorescence, but with respect to local differences in Raman activity. If the structure of the test specimen is known with sufficient accuracy, a traceable calibration of the Raman microscope becomes possible.

Under a test specimen is understood in particular a body whose structure is known exactly. For example, the structure can be measured with high precision and documented by a test certificate. A body which, although having a structure whose structure is not known with sufficient accuracy, is unsuitable as a test specimen. For a test specimen according to the invention for calibrating a Raman microscope, the structure, i. H. in particular, the expansions of the structural layer in the x and y directions with an expanded measurement uncertainty of 1% (expansion factor k = 2) known.

A Raman-active material is understood in particular to mean a material which has a polarizability that changes during a molecular or lattice vibration.

There is the possibility that the first material and thus the substrate is Raman-active, and the second material and thus the calibration structure is Raman-inactive, as well as the opposite case that the first material and thus the substrate is Ramaninaktiv and the second Material and thus the calibration structure is Raman active. All that matters is that the Raman image has a sufficiently high contrast between the substrate and the calibration structure. This may also be the case when both the first and second materials are Raman active. The decisive factor is the contrast that results after the addition of the signals of both materials at the wavenumber considered for calibration.

A calibration structure is understood to mean a structure which is shaped and dimensioned in such a way that it is suitable for the calibration of the microscope. A suitable structure for the calibration in one dimension has z. B. the shape of a tape measure incl. The associated scale. In other words, the calibration structure is suitable for measuring measured quantities such as distances, angles u. Ä. With known sizes, which are usually known due to a prior measurement of the specimen to compare. The calibration structure is generally largely or completely non-transparent to visible light and generally to excitation radiation radiated by the Raman microscope. SERS substrates often have similar structures, but are still not suitable for calibration because the dimensions are not in the required order of magnitude.

It has been found that a pronounced topography is disadvantageous for such a test body, since edge effects in the form of intensity peaks at the edges of the calibration structure can result. Therefore, it is necessary that the calibration structure is as thin as possible. A thin structure is understood to mean a structure whose thickness is less than 30 nm, preferably less than 20 nm. This results in a nearly topographieloser specimen.

It is desirable to have a high contrast, that is to say a large ratio between the Raman activity of the regions covered by the calibration structure and the Raman activity of the uncovered regions of the substrate. The contrast can also be caused by differences in the wavelength of the radiation emitted as Raman radiation.

The contrast can be influenced on the one hand by the material selection and on the other hand by the layer thickness of the calibration structure. A thicker layer produces better shading of the underlying substrate, thus resulting in lower Raman activity in the area of the calibration structure. At the same time, the edge thicknesses increase with the layer thickness as described above, so that a compromise between the desired high contrast and undesired edge effects must be found here. Advantageously, therefore, the calibration structure has a thickness between 10 and 25 nm, preferably between 15 and 20 nm. In tests carried out by the inventors, specimens with a layer thickness of the calibration structure between 15 and 20 nm have been found to be optimal.

An advantage of a test specimen according to the invention is that it allows the calibration of the adjustment of the microscope stage in point-by-point or line-by-line scanning based on the Raman emission. Furthermore, the point spread function of the optical resolution in the case of point-by-point or line-by-line scanning can be determined on the basis of the Raman signal.

The test specimen according to the invention has calibration structures which have no edge artifacts and are therefore virtually topographical. It is possible to produce a test specimen which has periodic structures which, with a small thickness of the calibration structure, have sufficient Raman contrast, even at small pitches, in order to calibrate small areas very precisely. Similarly, periodic structures may be present on the specimen that are sufficiently large to calibrate long ranges.

The different patterns of the calibration structure can be in different sizes on the Test specimens may be present to cover all common combinations of excitation wavelengths and objective as well as different pitches and different circle diameters.

By using two-dimensional, orthogonal structures, both major axes can be calibrated simultaneously and distortion in the image can be visualized. A return to the meter with reference analysis (AFM, REM) is possible. Furthermore, the test piece can be wetted with a drop of immersion liquid and cleaned again, since it can be carried out resistant to organic solvents.

A problem in the production of a test specimen according to the invention may be a poor adhesion of the calibration structure on the substrate. A development of the invention therefore provides that there is an adhesion promoter between the substrate and the calibration structure, wherein the thickness of a layer formed by the adhesion promoter is preferably less than 5 nm. Such a primer can, for. B. be titanium. In principle, however, any other material which has good adhesion to the substrate and on which the calibration structure can be applied is also suitable.

Semi-metals such as silicon and germanium are preferably conceivable as substrates, the Raman activity of germanium being significantly lower than that of silicon. Quartz is a suitable non-metallic material for a Raman active substrate. Raman-inactive substrates can consist of metals with the crystal systems fcc, hcp and bcc (gold, palladium, titanium or platinum). These metals are also suitable Raman-inactive materials for the structural layer and are capable of "shadowing" the underlying substrate to give a high contrast in the Raman image. Particularly suitable as the material of the calibration structure is an alloy of gold and palladium. The ratio of gold to palladium z. B. in the range between 40:60 and 70:30. In other words, the calibration structure contains 40 to 70% gold and 60 to 30% palladium. A suitable material for a Raman-active structure layer or calibration structure is in addition to silicon and germanium z. As graphene, consisting of only a monolayer modification of the carbon. So if z. For example, if a combination of graphene on silicon is selected, then the calibration structure will be seen in the Raman image at the wavenumber used for calibration, and the substrate will remain dark, whereas, for example, in FIG. Gold on silicon provides the inverse image. Graphene is further characterized by a very good adhesion even without adhesion promoter on the substrate.

To allow for full calibration, the calibration structure may include several of the following patterns: checkerboard pattern, dot grid, grating, dot scatter centers, large areas with sharp edges. Any combinations are possible depending on the specific requirements.

A pattern is any regular structure that allows calibration. As a rule, a pattern repeats after a certain spatial period. However, this is not absolutely necessary, so can a "large area with sharp edges" from any image of an object or z. B. consist of a character.

Two-dimensional, orthogonal structures are considered as "checkerboard patterns". These can be made in different sizes, but are at least 200 microns × 200 microns in size. Such checkerboard-like structures are suitable for simultaneously calibrating both major lateral axes without requiring rotation of the sample, ie the specimen. It has been found that calibrating a single major axis in the line mapping process can result in different table calibration factors than two-dimensional mapping. If the calibration structure has a checkerboard arrangement, this can continue to serve for the simultaneous assessment of the uniformity of the line shift, which would not be possible by means of a calibration structure which has only point grids. Such a two-dimensional structure need not necessarily be orthogonal and a checkerboard pattern, also conceivable z. B. diamond-like or hexagonal structures.

A "grating" is understood to mean a periodic, one-dimensional structure which may have different distances between the individual parts of the structure, so-called pitches. Of course, the structures are not one-dimensional in the strictly geometrical sense, but have a width of at least 200 μm. The geometric structure of the grating allows the measurement point recording over the entire adjustment range of the microscope stage used. The grids also allow 1D calibration at very small pitches on the order of near optical resolution due to the increased contrast to the two-dimensional structures.

The structures are advantageously each made several times with different pitches. Thus, a large pitch of z. B. 4 microns to order over a large calibration range at a pitch of z. B. 1 micron to ensure the Nyquist theorem minimum sampling. A relatively small pitch of 0.8 μm provides a sufficient number of periods for a high-resolution calibration of small areas with a pitch of z. B. 0.1 micron sure. Another pitch from Z. B. 2 microns makes it possible to cover different excitation wavelengths or lens combinations and step sizes.

As further features of the calibration structure, individual point scattering centers may be present. These can be z. B. diameter of 100 nm, 200 nm and / or 400 nm. With the help of such Punktstreuzentren then the detected point spread function and thus the optical resolution can be determined. There may also be duplicate point spread centers. The diameters can correspond to those of the individual Punktstreuzentren; Expediently, the center distance in the case of the 100 nm point scattering centers is four times (400 nm), otherwise double (400 nm, 800 nm) of the diameter. On the basis of such double Punktstreuzentren the optical resolution z. B. validate according to the Rayleigh criterion.

The point scattering centers can be used to determine the profile of the excitation radiation. The laser beam of a Raman microscope usually has a profile which substantially corresponds to a Gaussian curve, the radiation intensity as a function of the distance to the center of the optimally circular laser beam thus follows a Gaussian function whose maximum lies in the center of the laser beam. However, this theoretical Gaussian profile is influenced by the further optical components of the structure, so that the Gaussian profile can only represent an approximation for a real construction. With the help of existing on the test specimen scattering centers, the real beam profile can be determined and used for the development of the acquired image data. This leads to a significant increase in the accuracy of the process.

Large areas with straight, sharp edges allow the determination of the half-width of the detected point spread function. For this purpose, the edge can be scanned with linemapping method and the obtained intensity function as Gaussian error function (erf) be fitted. A "large area" is understood to mean a homogeneous, uninterrupted area which has an edge length which corresponds at least to five times one period of the regular structures used, which are eg. B. has an edge length of 5 microns × 5 microns. A "sharp edge" is understood to mean that the transition region between the structural material and the uncoated substrate is as small as possible, ideally equal to zero. Accordingly, all calibration structures used have sharp edges. Preferably, the layer thickness of the calibration structure increases over a distance of less than 50 nm from less than 5% to at least 95% of the nominal layer thickness. A straight edge is understood to mean a maximum deviation from a geometric straight line of 50 nm to 1 μm.

It is expedient to design the calibration structure redundantly on the substrate, so that preferably the structure comprises a plurality of different patterns and each pattern is present at least 10 times identically. It is less expensive to apply to a single test specimen a plurality of possibly identical structures, as to produce a plurality of test specimens, on each of which only a single structure is applied. As a result of the mentioned redundancy, it is thus possible with little effort to avoid the faulty production of a single structure rendering the entire test body unusable. Instead, it can be determined which structures were produced without errors, so that only these are used for calibration. It is then not necessary to use another specimen, but instead an identical structure applied in another area of the specimen is simply used for calibration.

An inventive test specimen can be produced by the following method: A silicon wafer as a substrate is z. B. coated with an electron beam sensitive paint or PMMA (polymethyl methacrylate) and then exposed by proximity corrected electron beam lithography. Under a photosensitive or electron beam sensitive paint is understood to mean any substance that is suitable for producing a mask, in addition to PMMA so z. As well as other electron-sensitive paints such as ZEP-520 Fa. Zeonrex Electronic Chemicals. The desired structures are transferred to the substrate. Subsequently, the exposed structure is developed. This gives a resist mask, in which the exposed structure has been transferred. On this resist mask, a thin layer of titanium and then a gold-palladium alloy is deposited in a thermal evaporator, wherein the alloy z. B. 60% gold and 40% palladium. As an alternative to a thermal resistance evaporator, an electron beam evaporator can also be used. It is important that the material is separated from a point source as possible.

Both the titanium layer and the working layer, ie the layer which later forms the calibration structure, are thin. Appropriately, of course, the titanium layer is thinner than the working layer, since from a certain thickness better adhesion is no longer achieved.

Possible materials for the calibration structure differ mainly in their fine granularity and in their behavior during the so-called lift-off, ie the removal of the resist mask. The Fine grain is important for producing high-precision, smallest structures and thus determines the smallest possible structure size and therefore indirectly also the accuracy of the calibration. It is also important that the material sticks well to the substrate during the lift-off, so that no faulty structures are produced. An alloy containing 60% by weight of gold and 40% by weight of palladium has been found to be advantageous. With such an alloy, structures can be made with sizes down to 100 nm at distances between the structures in the range down to 10 nm.

Subsequently, the resist mask is removed by a solvent. The exposed pattern is now transferred to the surface of the wafer as a metallic structure. The structures produced in this way, ie the entire material applied to the wafer material, then has a thickness of z. This is composed of a thin adhesion promoter layer, which consists of an approximately 3 nm thick titanium layer, and a working layer, which consists of gold and palladium and is about 20 nm thick.

The invention will be explained in more detail below with reference to the accompanying figures. Show it

1 a schematic representation of a test body according to the invention in plan view,

2 a schematic representation of a layout of a calibration structure,

3 an enlarged view of a single area of a calibration structure,

4 an enlarged view of another single area of a calibration, and

5 a schematic representation of a section through a test specimen according to the invention.

1 shows an embodiment of a test body according to the invention in plan view. The test piece 10 consists of a silicon chip as a substrate 12 as well as the applied calibration structures 14.1 . 14.2 , .... The chip has an edge length of about 3 cm × 3 cm. Other edge lengths of z. B. 2 cm × 2 cm, 4 cm × 4 cm but are just like rectangular substrates 12 , z. B. with an edge length of 3 cm × 4 cm, conceivable and suitable. Recognizable are the calibration structures shown as rectangles 14.1 . 14.2 , .... In total there are 20 identical calibration structures 14.1 . 14.2 , ... on the silicon chip 1 applied. A higher or lower number, eg. 5 × 5 = 25, 4 × 4 = 16 or 10 × 10 = 100 calibration structures 14.1 . 14.2 , ... but is also possible. Each of these calibration structures 14.1 . 14.2 , ... is arranged in a rectangular field, which has edge lengths of a = 4.5 mm and b = 3.5 mm. Such an area, in this case less than 20 mm ^2, is sufficient to accommodate all the necessary patterns. A test piece 10 on the one hand, the size of such a field would make it difficult to handle; on the other hand, with a substrate having an edge length of at least 2 cm, it would be possible to apply one drop of an immersion liquid, eg, B. of oil, applied to the specimen and also to remove. During cleaning, the test specimen can be collected and completely cleaned in a bath of organic solvents. Therefore, a larger chip becomes practical as a substrate 12 is selected. On this then a plurality of identical calibration structures 14.1 . 14.2 , ..., which greatly reduces the reject rate.

2 sketchy shows an enlarged view of a calibration structure 14.1 . 14.2 In one such calibration structure 14.1 . 14.2 , ... out 1 containing rectangle are so in the 2 12 substructures or patterns shown 18.1 . 18.2 , ... available. The patterns shown in column A are: CV Raman as a pattern with large areas and sharp edges, squares with sharp edges, single and double point scattering centers; in column B: Two-dimensional orthogonal structures in the form of checkerboard patterns, in column C: point gratings of different distances and in column D: one-dimensional structures in the form of line gratings.

The patterns are each available in different sizes. Thus, the diameters of the point scattering centers vary 18.2 . 18.3 as well as the distances of the points 18.4 . 18.5 . 18.6 , the lines 18.7 . 18.8 . 18.9 and the individual boxes of the checkerboard pattern 18:10 . 18:11 . 18:12 , In column B, the peridiodicities of the structures are indicated by way of example, which coincide here with the pitches. In the upper chessboard 18:10 the period is 4 μm, in the middle chessboard 18:11 2 μm and in the lower chessboard 18:12 0.8 μm. A single box of chess boards 18:10 . 18:11 . 18:12 So has an edge length that corresponds to half of the period.

3 shows in the lower area a single calibration field 20.1 as a section of a calibration structure 14.2 , which is shown schematically in the upper area again as an overview. The other calibration fields shown in the upper area 20.2 . 20.3 , ... can use other structures, such as those in 2 are shown included. The in the calibration field 20.1 contained circular and square Punktstreuzentren 18.2 are available in different dimensions, the edge lengths a vary from 0.1 μm (below) to 3.2 μm (above). The edge length l of the calibration field 20.1 is 400 μm × 400 μm.

4 shows in the lower area a single calibration field 20.2 as a section of a calibration structure 14.2 , which is shown schematically in the upper area again as an overview. The other calibration fields shown in the upper area 20.1 . 20.3 , ... can use other structures, such as those in 2 are shown included. The illustrated calibration field 20.2 has a checkered structure 18:11 As a two-dimensional, orthogonal structure and is suitable for calibrating the lateral major axes of the sample. The period a is in the case shown 0.8 microns, but can also z. B. 2 microns or 4 microns

5 shows a section through a region of a test body according to the invention 10 , The representation is not to scale. It can be seen that the substrate made of silicon 12 makes up the largest proportion of the test specimen. Area wise is the silicon substrate 12 not covered by another layer, so that the incident from above excitation radiation 28 Raman microscope can penetrate unhindered to the substrate. The high Raman activity of silicon at the corresponding sites is determined by Raman scattering 30 measured. In the picture left and right are a thin layer 16 the thickness d ₂ of titanium and a layer 22 the thickness d _{1 applied} a gold-palladium alloy. Through these layers 16 . 22 becomes the silicon substrate 12 covered, so that neither excitation radiation 28 unhindered to the substrate 12 nor any of the resulting Stokes line photons of attenuated Raman scattering 32 can easily go back to the microscope objective and from there to the spectrometer. The edges 24 . 26 are perpendicular to a good approximation. A highly accurate design of the structure is necessary to allow accurate calibration.

The cover layer or working layer 22 of gold and palladium is Raman inactive, so there is a high contrast in the Raman image between the areas that are coated and the uncoated areas. The areas coated with titanium and the gold-palladium alloy form the calibration structure 14 , The total thickness of the calibration structure 14 is the sum of the two thicknesses d _{1 of} the gold-palladium alloy and the thickness d _{2 of} the titanium.

LIST OF REFERENCE NUMBERS

10

specimen

12

substratum

14

calibration structure

16

bonding agent

18

template

20

calibration field

22

work shift

24

edge

26

edge

28

incident radiation

30

Raman scattering

32

weakened Raman scattering

a

Diameter, period

d ₁

Thickness of the working layer

d ₂

Thickness of the adhesion promoter layer

l

edge length

Claims (10)

Test specimen ( 10 ) for calibrating a Raman microscope with a substrate ( 12 ) of a first material, and a calibration structure applied thereto, ( 14 ) of a second material different from the first material, wherein at least one of the first material and the second material is Raman active, characterized in that the calibration structure ( 14 ) has a thickness d ₁ of less than 30 nm Test specimen ( 10 ) according to claim 1, characterized in that the potash structure ( 14 ) has a thickness d ₁ between 10 and 25 nm, preferably between 15 and 20 nm. Test specimen ( 10 ) according to claim 1 or 2, characterized in that between the substrate ( 12 ) and the calibration structure ( 14 ) a bonding agent ( 16 ), wherein the thickness d _{2 of} one of the adhesion promoter ( 16 ) is preferably less than 5 nm. Test specimen ( 10 ) according to claim 1 or 2, characterized in that the calibration structure ( 14 ) consists of a Raman-active material. Test specimen ( 10 ) according to one of claims 1 to 3, characterized in that the calibration structure ( 14 ) consists of 40 to 70 wt .-% of gold and 60 to 30 wt .-% of palladium. Test specimen ( 10 ) according to one of the preceding claims, characterized in that the calibration structure comprises a plurality of point scattering centers. Test specimen ( 10 ) according to one of the preceding claims, characterized in that the calibration structure ( 14 ) includes several of the following patterns: checkerboard pattern, dot grid, grating, dot scattering centers, large areas with sharp edges. Test specimen ( 10 ) according to one of the preceding claims, characterized in that the calibration structure ( 14 ) on the substrate ( 12 ) is redundant, wherein preferably the calibration structure ( 14 ) a plurality of different ones Includes patterns and each pattern is at least 10 times identical. Method for producing a test specimen ( 10 ) for calibrating a Raman microscope with the steps: a. Applying a layer of a photosensitive or electron beam-sensitive lacquer to a substrate ( 12 ), which consists of a first material b. Exposing the lacquer layer with a suitable pattern for the calibration of a Raman microscope by means of electron beam lithography c. Depositing a layer of thickness d ₁ of less than 30 nm on the substrate of a second material different from the first material, wherein the first material and / or the second material is Raman-active d. Removing the layer of electron beam-sensitive paint Use of a test specimen ( 10 ) according to any one of claims 1-8 for calibration of a Raman microscope.

Images