Method for non-destructive authenticity testing of an object

Abstract

The invention relates to a method for non-destructive authenticity testing of an object by means of an X-ray fluorescence analysis, in which X-ray radiation (16) is directed onto the object (24), the X-ray fluorescence (31) emitted by the object (24) is detected and delivered to an evaluation device (28), and a measurement spectrum (41) of the object (24) is determined, wherein at least one model MV relating to an uncoated sample composed of a main body, and at least one model MB relating to a coated sample consisting of a main body with at least one coating, are stored in the evaluation device (28), and a spectrum of the model MV and a spectrum of the model MB which come closest to the detected measurement spectrum (41) are selected and varied, and whichever of the models MB and MV corresponds more closely to the detected measurement spectrum (41) is output as the result.

Description

[0001] MM 60112

[0002] Methods for non-destructive authentication of an object

[0003] The invention relates to a method for non-destructive authentication of an object using X-ray fluorescence analysis.

[0004] From EP 2 623 929 Bl, an authenticity test for gold jewelry using X-ray fluorescence is known. This method is used to detect counterfeit gold jewelry. The method can determine whether the gold jewelry is made of a gold alloy or whether it is merely gold-plated. In this method, the gold jewelry is excited with X-rays, and the intensity of a first energy, corresponding to an L-alpha energy emission line of gold, is measured. Furthermore, the intensity of a second energy, corresponding to an L-beta energy emission line of gold, is measured. A ratio is calculated from these two measured values. If the calculated ratio is greater than a first predetermined value or less than a second predetermined value, the statement is made that the gold jewelry is gold-plated.

[0005] The invention is based on the objective of proposing a method for the non-destructive authentication of an object in order to determine whether the object has at least one coating on a base body of the object or whether the object is made of an alloy free of any coating.

[0006] This problem is solved by a method in which at least one model MB, consisting of at least one spectrum with at least one fluorescence line from a preferably specified sample, is stored in the evaluation unit, and at least one model MV, which comprises an uncoated

[0007] MM 60112p0.docx Oct 15, 2025 09:33

[0008] 4449 / hu / bn / si / hu MM 60112

[0009] Page 2 of 17

[0010] The model relates to a sample consisting of a base body, preferably made of a solid material, such as an alloy with alloying elements, wherein at least one spectrum with fluorescence lines of the uncoated sample is stored in the model MV, and wherein a predetermined selection of chemical elements is assigned to the uncoated sample, and wherein at least one model MB is stored in the evaluation device, which relates to a coated sample consisting of a base body with at least one coating, wherein at least one spectrum with fluorescence lines of the coated sample is stored in the model MB, and wherein at least the predetermined selection of chemical elements of the uncoated sample is stored for the coated sample.and wherein one chemical element from the predetermined selection of chemical elements is assigned to the coating and the other chemical elements from the predetermined selection of chemical elements are assigned to the base body, and wherein the spectrum of model MV and the spectrum of model MB that most closely matches the acquired measurement spectrum is selected, and the selected spectrum of model MV and model MB is varied and stored as the varied spectrum of model MV and as the varied spectrum of model MB, and wherein the varied spectrum of model MV and the varied spectrum of model MB are compared with the acquired measurement spectrum of the object, and the result is the output of model MB or model MV that shows a greater match with the acquired measurement spectrum.

[0011] This method has the advantage that, through at least one model MB and at least one model MV stored in the evaluation unit, at least one algorithm can be used to vary models MB and MV. By comparing these variations with the recorded measurement spectrum, a statement can be made as to whether the object is coated, made of a solid material such as an alloy, or uncoated. This method allows the object's structure to be determined non-destructively. Furthermore, this method has the advantage that it can be performed without prior knowledge of the object under investigation, i.e., whether it has a layered structure.

[0012] MM 60112p0.docx MM 60112

[0013] Page 3 of 17: Whether the object is made of a solid material or exhibits such properties, a non-destructive analysis can be performed and a conclusion can be drawn about the object being examined. This method is not limited to assessing gold jewelry, whether it consists of an alloy or is gold-plated, but can be used for various chemical elements.

[0014] Preferably, a first quantitative value is determined between the varied spectrum of model MB and the recorded measurement spectrum, and a second quantitative value is determined between the varied spectrum of model MV and the recorded measurement spectrum. Depending on the selected quantitative value, either model MB or model MV is output as the result, exhibiting the larger or smaller quantitative value, respectively. This allows for simple computational evaluation.

[0015] The quantitative measure is a measure of the goodness of fit, also called the fit of the model, between the measured and modeled spectrum. The measure of goodness of fit allows a statement about the discrepancy between the theoretical values ​​of the investigated random variables, which are expected or predicted by the model, and the values ​​actually measured.

[0016] A goodness of fit can be measured, for example, by an RMS value. Alternatively, a mean square error (MSE) or a chi-square test (x) can also be used. 2 - test), a correlation coefficient (R 2 ), a Root Mean Squared Error (RMSE) or a Normalized Mean Absolute Error (NMAE) can be selected as the goodness of fit.

[0017] When using the rms value for the goodness of fit, the following formula is used, for example: rms = 1 - * 100, where the parameter a is determined from a = I s(k) - m( / c)| and the parameter b is determined from b= b = m( / c)| . Here, s(K) represents the measurement spectrum and m(K) represents the spectrum of the respective model. K is the

[0018] MM 60112p0.docx MM 60112

[0019] Page 4 of 17: Running index for the respective channel within a spectrum under consideration. N represents the number of channels. Typically, n is assumed to be 1024 channels.

[0020] Model MB applies to at least one coating on a substrate. Model MB thus consists of I-layers. Layer I can represent a coating on a substrate, or it can represent the substrate itself. Therefore, in model MB, the number of I-layers is at least 2. It is preferably assumed that the coating has a finite thickness and is applied to a substrate made of solid material. The substrate is defined as having an infinitely thick layer if increasing the thickness of this layer does not result in any measurable difference.

[0021] Model MV applies to the base body made of a solid material, i.e., without at least one coating. This model is structured such that, analogous to model MB, this single layer, or the base body itself, is assumed to be infinitely thick.

[0022] When selecting this quantitative measure in the form of an rms value, the model MB or MV is chosen as the result for the performed authenticity check, where the quantitative measure indicates the greatest match between the measurement spectrum and the model spectrum.

[0023] The goodness of fit does not necessarily have to be calculated over the entire measurement spectrum s(K) or model spectrum m(K). The index K can also be a subset of the total spectrum. Specifically, the goodness of fit does not have to be calculated over the entire spectrum (e.g., 1024 channels). In some cases, a well-chosen, physically sound selection of, for example, 500 channels may suffice to enable an analysis regarding the coating on a substrate or an uncoated substrate.

[0024] This enables a standardized evaluation.

[0025] MM 60112p0.docx MM 60112

[0026] Page 5 of 17

[0027] Preferably, the selected spectrum of model MB and model MV, which most closely approximates the acquired measurement spectrum, is varied by an algorithm. In particular, the use of a Levenberg-Marquardt optimization algorithm may be provided.

[0028] Another advantageous embodiment of the method provides that the same number of chemical elements, which are at least contained in the object, are stored for both model MV and model MB. Thus, not only are the same chemical elements used as the basis for the models, but also the same quantity. This improves the evaluation and the accuracy of the prediction regarding the type of object, specifically whether it is coated or uncoated.

[0029] For example, it may be possible to select at least one specific fluorescence line of the object for recording its measurement spectrum. For instance, L-fluorescence lines specific to gold could be selected. Furthermore, K-fluorescence lines, representing a base material of the object such as iron, nickel, copper, or similar materials, could also be selected. This allows the method to be better tailored to the specific object being tested.

[0030] Advantageously, the system does not base the measurement spectrum on individual, predefined fluorescence lines and their intensities, but rather on capturing the entire X-ray fluorescence spectrum of the object and using it to verify its authenticity. Using the entire X-ray fluorescence spectrum as the measurement spectrum offers the advantage of allowing for the evaluation of significantly more information about the object than is possible when selecting individual fluorescence lines.

[0031] This makes it possible to detect additional fluorescence lines from existing chemical elements in the object's base material, relative to each other, in a coated object. Furthermore, the detection of the entire

[0032] MM 60112p0.docx MM 60112

[0033] Page 6 of 17

[0034] The X-ray fluorescence spectrum allows the method to be applied to a wide variety of material systems.

[0035] The invention, as well as further advantageous embodiments and developments thereof, are described and explained in more detail below with reference to the examples shown in the drawings. The features that can be derived from the description and the drawings can be applied individually or in any combination according to the invention. The drawings show:

[0036] Figure 1 shows a schematic view of an X-ray fluorescence measuring device.

[0037] Figure 2 shows a schematic representation of a model MV and a model MB.

[0038] Figure 3 shows a flowchart for performing non-destructive authentication of an object and outputting a result.

[0039] Figure 4 shows a schematic view of a diagram for a varied model MB of an object made of an alloy.

[0040] Figure 5 shows a schematic view of a diagram for a varied model MV of an object made of an alloy,

[0041] Figure 6 shows a schematic view of a diagram for a varied model MB of an object with a base body and a coating, and

[0042] Figure 7 shows a schematic view of a diagram for a varied model MV of an object with a base body and a coating.

[0043] MM 60112p0.docx MM 60112

[0044] Page 7 of 17

[0045] Figure 1 schematically and exemplarily depicts an X-ray fluorescence measuring device 11 for X-ray fluorescence analysis. This X-ray fluorescence measuring device 11 comprises an X-ray tube 13 with a heated cathode 14 as the X-ray source 12, from which electrons are emitted and directed towards an anode 15. These electrons are accelerated towards the anode 15 by an accelerating voltage UB. There, the electrodes are decelerated and generate X-ray radiation 16. A shutter 17 can be provided on the X-ray tube 13, which can be controlled to allow the X-ray radiation 16 to escape.

[0046] An X-ray optic 18, which may be, for example, a mono- or polycapillary lens, is preferably positioned in the X-ray beam 16. Alternatively or additionally, a collimator 19 may also be provided. Furthermore, a mirror 21, which is transparent to the X-ray beam 16, is provided in the beam path of the X-ray beam 16. This allows an image to be captured by a camera 22 from a measuring point 23 on an object 24. Preferably, an imaging optic 27 is provided between the mirror 21 and the camera 22. This camera 22 is preferably connected to an evaluation device 28, such as a monitor 29.

[0047] According to the illustrated embodiment, object 24 can have a coating 25 on a base body 26. Multiple coatings 25 can also be applied to the base body 26. Alternatively, object 24 can also consist of an uncoated base body. In other words, object 24 then consists of a solid material, in particular an alloy, for example of different metals, and is free of any coating.

[0048] At measuring point 23, X-ray radiation 16 generates X-ray fluorescence radiation 31, which is detected by at least one detector 32. A measurement spectrum 41 detected by the detector 32 can, for example, be displayed as a graph 32 on a screen. By evaluating a detected measurement spectrum 41 or a measured energy spectrum of the X-ray fluorescence radiation 31,

[0049] MM 60112p0.docx MM 60112

[0050] Page 8 of 17. The material composition of object 24 and / or the coating 25 and / or the layer thickness of the layers and / or the layer system can be determined in a manner known per se. In the present case, it should be determined in a simple manner, without knowing the structure and composition of object 24, whether object 24 comprises at least one coating 25 on the base body 26, or whether object 24 consists only of the base body 26.

[0051] For the non-destructive authentication of object 24 using the X-ray fluorescence measuring device 11, at least one model MB and at least one model MV are stored in the evaluation unit 28. These models MB and MV are used for analysis and evaluation in order to determine whether the object is a forgery or whether it possesses the required properties.

[0052] The MV model (model for a solid sample) consists of only one layer, n = 1. This layer, n = 1, is formed by the base material. The base material can be an alloy containing several chemical elements. The MB model (model for a coated sample) consists of layers, n = 2, where at least n = 2. This means that the MB model comprises a base material (layer n = 1) and at least one coating (layer n = 2, 3, ...) on the base material.

[0053] The term "model" can refer to an abstract physical model of the interaction between X-rays and matter. Matter here refers to the individual atoms of different chemical elements. In the model, atoms of chemical elements are assigned to different layers, if different layers exist. The model parameters are the absolute and relative number of these atoms in each layer.

[0054] Preferably, the model is formed from a spectrum with at least one fluorescence peak from a sample or a predefined sample. For model MV, this predefined sample consists of a known solid material, for example, a known alloy with

[0055] MM 60112p0.docx MM 60112

[0056] Page 9 of 17

[0057] Alloy component. This specified sample is free of any coating. For example, the sample could be a gold alloy with a specified carat weight. This MV model consists of only a single layer.

[0058] The specified specimen for model MB consists of a base body made of a known solid material or a known alloy with an alloying element, and at least one known coating applied to the base body. Preferably, a specified specimen for model MB is produced in which the components of the coating are not incorporated into the base body. This application is particularly useful for verifying the authenticity of gold jewelry.

[0059] Figure 2 shows a schematic view of model MV and model MB, including their chemical elements. For the base body according to model MV, a predetermined number of chemical elements present in the solid material are used as a basis. The chemical elements contained in the solid material can form the basis for selecting the chemical elements used for model MV. The number of chemical elements for model MV can be selected. For example, the base body consists of an alloy containing a variety of chemical elements, and three chemical elements are selected and used for model MV. These could be, for example, gold (Au), silver (Ag), and copper (Cu).

[0060] For model MB, at least the same chemical elements are used as for model MV. Thus, model MB includes gold (Au), silver (Ag), and copper (Cu) as chemical elements. For example, if it needs to be determined whether object 24 consists of gold or a gold-coated base, model MB will assign the chemical element gold (Au) to the coating, while assigning the remaining chemical elements to the base. The selection of chemical elements and / or the number of chemical elements...

[0061] MM 60112p0.docx MM 60112

[0062] Page 10 of 17, which form the basis for the respective models MV and MB, are only examples. However, it is intended that the selected chemical elements for model MV will also form the basis for model MB, with one of the chemical elements being assigned to layer n = 2, i.e., the coating. If the object to be tested is an item comprising a base body with two coatings, each coating will be assigned a chemical element, with these chemical elements assigned to the coatings differing from one another.

[0063] Preferably, a large number of MB and MV models are stored in an evaluation unit. The number of models is limited by the number of chemical elements listed in the periodic table.

[0064] Figure 3 shows a schematic diagram illustrating the procedure for the non-destructive authentication of object 24. Such a procedure is particularly useful for identifying counterfeits. For example, a dealer in jewelry or similar items might claim that the piece of jewelry for sale is made entirely of a gold alloy, when in reality it consists only of a gold alloy layer a few millimeters thick on a precious base material, such as brass.

[0065] The object 24 is positioned in the X-ray fluorescence measuring device 11 and irradiated with X-ray radiation 16. A measurement spectrum 41 of the object 24 is acquired according to step 40 in Figure 2. The acquired measurement spectrum 41 can consist of individual, or even selected, fluorescence lines. Advantageously, the acquired measurement spectrum 41 comprises the entire X-ray fluorescence spectrum acquired by the object 24. This complete X-ray fluorescence spectrum can also include diffuse X-ray scattering.

[0066] After the measurement spectrum 41 has been acquired, at least one model MB and at least one model MV are selected in the evaluation unit 28. This can be done according to step 42. The stored model MB and model MV that most closely match the measurement spectrum 41 are selected.

[0067] MM 60112p0.docx MM 60112

[0068] Page 11 of 17

[0069] Starting with the models selected in step 42, step 43 involves varying each model until it most closely matches the measurement spectrum 41. Advantageously, this model variation is performed using an iterative fitting algorithm. The model parameters MB and MV are varied. These parameters can be based on the relative and absolute number of atoms in the layers of the respective model. A stepwise variation of the models is preferably performed using a Levenberg-Marquardt optimization algorithm. This variation of the models MB and MV continues until the spectra of the varied models MB and MV have been fully varied, meaning that a further change in a quantitative parameter of the respective model is no longer computationally possible or significant, and the spectra fall below a defined value compared to the measurement spectrum of object 24.as a quantitative measure. This quantitative measure is a so-called rms value, which is determined as described previously.

[0070] In step 44, the quantitative value of model MB and the quantitative value of model MV are compared. The quantitative value that deviates least from the recorded measurement spectrum 41 is output as the result. If, according to step 44, the quantitative value of model MB is closest to the measurement spectrum 41 of object 24, this model MB is selected and output as the result of the authentication check in step 45. This means that sample 24 is a base body 25 with at least one coating 26. If, according to step 44, the quantitative value for model MV is closest to the recorded measurement spectrum 41, and model MV is output in step 45, and sample 24 is a solid material or an alloy without a coating, then the result is...

[0071] Figures 4 and 5 illustrate the previously described sequence of individual steps, particularly when comparing the varied models MB and MV with the recorded measurement spectrum 41. The example shown in Figures 4 and 5 is based on an object 24 which

[0072] MM 60112p0.docx MM 60112

[0073] Page 12 of 17 consists of an alloy or a solid material and, for example, includes gold as a component. The authentication test is then started as described in Figure 2. The diagram in Figure 3 then shows the recorded measurement spectrum 41. The two arrows 48 point to the L-fluorescence lines specific to gold. Furthermore, to determine the actual structure of object 24, model MB was selected and varied using the iterative fit algorithm. The resulting spectrum is shown as line 47. A comparison of line 41, which represents the recorded measurement spectrum 41, and line 47, which corresponds to the varied model MB, reveals a discrepancy. This discrepancy can be determined by the quantitative value, in particular the rms value. In this example, an rms value of rms = 0.08 was determined.

[0074] Figure 5 again shows the recorded measurement spectrum 41 as in Figure 4. The model MV is then selected and varied using the iterative fitting algorithm. This results in a line 49 for the spectrum of the varied model MV, which corresponds almost exactly to the line of the measurement spectrum 41. The resulting quantitative value, or rms value, is, for example, rms = 0.008.

[0075] A comparison of the quantitative size of model MV and model MB shows that the agreement between the varied model MV and the recorded measurement spectrum 41 is significantly better than that of the varied model MB. In other words, the deviation of the quantitative size of model MV from the measurement spectrum 41 of object 24 is smaller than the deviation of the quantitative size of model MB. Therefore, the result is that object 24 is made of a solid material or an alloy.

[0076] Figures 6 and 7 show a further embodiment. This further embodiment starts with an object 24 in which a coating 25, for example a gold coating, is applied to a base body 26, for example steel. The measurement spectrum 41 acquired by the object 24 is shown in the diagram according to

[0077] MM 60112p0.docx MM 60112

[0078] Page 13 of 17

[0079] Figure 6 illustrates this. Due to the advantageous acquisition of the entire X-ray fluorescence spectrum, arrows 48, for example, indicate the gold L-fluorescence lines. Furthermore, arrows 51 point to K-fluorescence lines. These are specific fluorescence lines for steel, for example, for the constituents iron and chromium. Now, the model MV is selected and varied using the iterative fitting algorithm. This results in line 52 shown in Figure 6. In determining the quantitative quantity, an rms value of rms = 0.08 is obtained in this example.

[0080] In the diagram according to Figure 7, the recorded measurement spectrum 41 of object 24 is again shown in analogy to Figure 6. Additionally, the spectrum of model MB was varied, in particular by the iterative fitting algorithm. This resulted in line 53 shown in Figure 7. Especially in the region of the L and K fluorescence lines, a significantly better agreement can be seen between the varied spectrum of model MB according to line 52 and the recorded measurement spectrum 41 than in Figure 6. For example, a quantitative rms value of rms = 0.015 is determined for model MB.

[0081] A comparison of the quantitative size of model MB and model MV determined under each assumption shows that in the present case, according to Figure 7, the quantitative size of model MB is smaller than the size of model MV, leading to the conclusion that the object in question is an object 24 with at least one coating 25 on a base body 26.

[0082] The preceding embodiments according to Figures 4 to 7 were discussed using gold L-fluorescence lines as indicators only as examples. Other and / or additional fluorescence lines can also be used. In particular, the entire X-ray fluorescence spectrum is considered.

[0083] MM 60112p0.docx

Claims

MM 60112 Page 14 of 17 Claims 1. Method for non-destructive authentication of an object using X-ray fluorescence analysis, - in which an X-ray radiation (16) is emitted from an X-ray source (12) and directed at the object (24), - wherein at least one detector (32) detects the X-ray fluorescence radiation (31) emitted by the object (24) and supplies it to an evaluation device (28) and determines a measurement spectrum (41) of the object (24) which includes at least one fluorescence peak, characterized in that - that at least one model MV is stored in the evaluation device (28) which relates to an uncoated sample from a base body, wherein at least one spectrum with fluorescence lines of the uncoated sample is stored in the model MV, and wherein a predetermined selection of chemical elements is assigned to the uncoated sample, - that at least one model MB is stored in the evaluation unit (28) which relates to a coated sample from a base body with at least one coating, wherein at least one spectrum with fluorescence lines of the coated sample is stored in the model MB, wherein at least the predetermined selection of chemical elements of the uncoated sample is stored for the coated sample, and wherein one chemical element from the predetermined selection of chemical elements is assigned to the coating and the other chemical elements from the predetermined selection of chemical elements are assigned to the base body, - that the spectrum of model MV and the spectrum of model MB, which is closest to the measured spectrum (41), MM 60112p0.docx MM 60112 Page 15 of 17 is selected and the selected spectrum of model MV and model MB is varied and saved as a varied spectrum of model MV and as a varied spectrum of model MB, - that the varied spectrum of model MV and the varied spectrum of model MB are compared with the recorded measurement spectrum (41) of the object (24) and the result is the model MB or the model MV which shows a greater agreement with the recorded measurement spectrum (41).

2. Method according to claim 1, characterized in that a first quantitative quantity is determined by comparing the varied spectrum of model MB and the recorded measurement spectrum (41) and a second quantitative quantity is determined between the varied spectrum of model MV and the recorded measurement spectrum (41), and that the larger quantitative quantity is output as the result.

3. Method according to claim 2, characterized in that an rms value according to the formula rms = (1 * 100 is determined, where the parameter a is determined from a = | s(k) - m( / c)| and the parameter b is determined from b = fc=ol m ( , where s(K) represents the measurement spectrum and m(K) represents the spectrum of the respective model.

4. Method according to claim 2, characterized in that the quantitative measure is a mean square error (MSE), a chi-square test (x 2 -test), a correlation coefficient (R 2 ), a Root Mean Squared Error (RMSE) or a Normalized Mean Absolute Error (NMAE) is selected as the goodness of fit.

5. Method according to one of the preceding claims, characterized in that the selected spectrum of model MB and model MV, which is closest to the detected measurement spectrum (41), is varied by an algorithm, in particular a Levenberg-Marquardt optimization algorithm. MM 60112p0.docx MM 60112 Page 16 of 17 6. Method according to one of the preceding claims, characterized in that the same number of chemical elements are stored in the model MV and the model MB, which are contained at least in the object.

7. Method according to one of the preceding claims, characterized in that the at least one specific fluorescence line for the material of the object (24) to be tested is selected as the measured measurement spectrum (41).

8. Method according to one of the preceding claims, characterized in that the entire X-ray fluorescence spectrum of the object (24) is evaluated as the measured measurement spectrum (41) of the object (24). MM 60112p0.docx

Images