Method for the determination of a crosslinking degree

Abstract

The invention relates to a method for the determination of the crosslinking degree of a polymer (1) in a sample (2), the method comprising the following steps: - exposing the polymer (1) to near infrared radiation (3) originating from a radiation source (4), in particular such that the near infrared radiation (3) passes the polymer (1) at least once, - determining a set of measurement values being indicative of the absorption of the near infrared radiation (3) by the polymer (1) in a measurement wavelength range, wherein the wavelength range is between 800 nm and 3000 nm, - feeding the set of measurement values to a multivariate data model, which multivariate data model is based on a multiplicity of sets of reference values, - obtaining, from the multivariate data model, an output value being indicative of the crosslinking degree of the polymer (1). The invention also relates to a device.

Description

[0001] Method for the determination of a crosslinking degree

[0002] The present invention relates to a method for the determination of the crosslinking degree of a polymer as well as to an apparatus.

[0003] Peroxidic crosslinked polymers are used in many fields of industry, such as in the production of photovoltaic modules as encapsulants for solar cells. Most often, elastomeric ethylene vinyl acetate is employed for this purpose, but also other peroxidic crosslinked polymers could be used. One decisive factor, which determines the quality of such polymers, is their crosslinking degree.

[0004] In industrial practice, one of the most important process parameter is the lamination time. In case of the production of photovoltaic modules, the lamination of the solar cells thus is the time-limiting step in the production process.

[0005] High productivity typically requires a short lamination time, which in turn leads to low crosslinking degrees. However, high product quality can only be obtained if the crosslinking degree is high, which requires a long lamination time. A conflict of interest arises from this, as a sufficient degree of crosslinking needs to be ensured, while the lamination time should be kept to a minimum.

[0006] It is therefore known to measure the crosslinking grade of peroxidic crosslinked polymers by analytical methods. The current standard for this is the determination of the so-called “gel content”, which is obtained after solvent extraction of a polymer sample. This approach is simple, but it has a number of disadvantages. Some of them are: The measurement is destructive and requires sampling, such that assembled products cannot be tested; the typical duration required for performing the measurement is between 16 and 24 hours, such that real-time measurement is not feasible; the measurement results may vary considerably depending on the method and parameters of the analysis, such that a comparison between different measurement values can be problematic.

[0007] Other analytical methods are known for the determination of the crosslinking degree, such as Raman spectroscopy. This method can be performed without taking physical samples and it is non-destructive. Therefore, it can be used for finished products, potentially even as an in-line method. However, some drawbacks remain, which hamper the widespread use of this technique in the present field of interest, in particular in relation to the instrumental complexity of such measurement devices.

[0008] An object of the present invention may be therefore be seen in overcoming the above- mentioned problems. A particular object may be seen in providing a method, which is non-destructive, reliable, and can be implemented with relatively low instrumental effort. Another object may be seen in providing a method, which can be used on finished photovoltaic products, even if the polymers to be measured are covered by one or more other layers of material.

[0009] It has now been surprisingly found that near infrared spectroscopy may be a solution for at least one of the above-mentioned objects, in particular if used in conjunction with multivariate data evaluation approaches. In particular, low instrumental complexity may be combined with high data reliability by using this method. In particular, a method for the determination of the crosslinking degree of a peroxidically crosslinked polymer is therefore described, wherein a polymer material is exposed to near infrared radiation, and wherein the measured absorption or transmission of the near infrared radiation is used for the determination of the crosslinking degree of the polymer.

[0010] The measurement of “absorption” and / or “transmission” also includes the possibility of a measurement by means of reflection, such as in particular by transflection.

[0011] It may be particularly provided that the determination of the crosslinking degree is based on or performed by multivariate data analysis. In detail, the measurement of the polymer may deliver a set of measurement values, in which each single data point, i.e., each measurement value, is indicative of the absorption of the near infrared radiation at a specific wavelength or a specific wavelength range of the polymer. The set of measurement values may comprise a multitude of data points, such as more than 500, more than 1 ,000, or even more than 5,000, data points. The set of measurement values may contain a number of independent variables, relating to wavelengths, along with responses to each independent variable. The set of measurement values may be visualized for example in the form of a transmission or absorption spectrum.

[0012] Classification and / or quantification of the crosslinking degree of a measured polymer of a sample may particularly be obtained by feeding the set of measurement values into a multivariate data model. The multivariate data model may be based on a multitude of sets of reference values, wherein the sets of reference values are obtained from the measurement of samples, which have a known crosslinking degree.

[0013] For generation of the multivariate data model, at least 3, or at least 4, or even at least 5, sets of reference values may be used.

[0014] In the sets of reference values, each single data point, i.e., each reference value, is indicative of the absorption of the near infrared radiation at a specific wavelength or a specific wavelength range of a reference polymer. Each set of reference values may comprise a multitude of data points, such as more than 500, more than 1 ,000, or even more than 5,000, data points. Each set of reference values may contain a number of independent variables, relating to wavelengths, along with responses to each independent variable. The sets of reference values may be visualized for example in the form of transmission or absorption spectra.

[0015] The multivariate data model may by a classification model or a quantification model, such as a principal component analysis (PCA)-model, for classification, or a principal component regression (PCR)-model or a partial least squares (PLS)-model, for quantification. Such models are well-known in the art and generally spoken, they provide for the identification of high-variance components from a set of values, while filtering out low-variance components. From a mathematical point of view, the multivariate data processing may project the sets of measurement and reference values into new coordinate systems, in order to reduce dimensionality of the data set.

[0016] The set of measurement values and the sets of reference values may contain the same or essentially the same number of data points. The wavelength range of the set of measurement values and the wavelength range of the sets of reference values may be the same or essentially the same.

[0017] The measurement wavelength range, i.e., the wavelength range in which the values of the sets of measurement and reference values lie, may be between 800 nm and 3,000 nm. This complete wavelength range may be the measurement wavelength range, or one or more specific wavelength ranges may be selected from this broad range.

[0018] The polymer, from which the set of measurement values is obtained, may be a sample or it may be part of a sample. The sample may also contain other materials apart from the polymer. In a specific embodiment, the polymer may be provided in the form of a layer of the sample.

[0019] The sets of measurement and reference values may be obtained by transmission or by transflection measurement. When performing a transmission measurement, the near infrared radiation passes the sample once, wherein the polymer is placed between the radiation source and the detector. When performing a transflection measurement, the near infrared radiation passes the sample twice, wherein the polymer is placed between the radiation source and a reflector, and wherein the detector is placed on the side of the radiation source. The reflector may be a part of the measurement device, but it may also be a part of the sample. In general, the reflector should be able to at least partially reflect the near infrared radiation.

[0020] In the sample, the polymer may be covered with a cover material or cover layer, which is at least partially transparent for the used near infrared radiation. The transmission grade of the cover material for the near infrared radiation in the measurement wavelength range may for example be at least 50%, or at least 80%. Such cover material may be glass or a polymer, wherein the polymer particularly is not a peroxidically crosslinked polymer.

[0021] The sample may particularly be a photovoltaic module, which comprises in the following order: a photovoltaic layer; a polymer layer; a cover layer. The photovoltaic layer may be formed by or comprise crystalline silicon or other semi-conductive material, which provides for a photoelectric effect. The polymer layer may be formed by or comprise the polymer of interest, namely a peroxidically crosslinked polymer. The cover layer may be formed by or comprise a glass layer. The thickness of the cover layer may be less than 10 mm, in particular less than 5 mm.

[0022] The sets of reference values may be obtained from the measurement of reference samples. The reference samples particularly have the same or essentially the same structure as the sample. This means for example that if the sample is a compound material of a crystalline silicon material, covered by an ethylene vinyl acetate (EVA) layer and provided with a glass cover layer, the reference samples should also be compound materials of a crystalline silicon material, covered by an ethylene vinyl acetate (EVA) layer and provided with a glass cover layer.

[0023] It may specifically provided that in the sample and in the reference samples, one or more of the following parameters may be the same or essentially the same: A thickness of the cover layer; a material composition of the cover layer; a thickness of the polymer layer; a material composition of the polymer layer. The crosslinking degree of the reference samples may be determined by solvent extraction of the polymer layer, such as by Soxhlet extraction. In particular, nonadhesive films may be provided between the polymer layer and the other parts of the reference samples, in order to allow for reliable removal of the polymer material. These non-adhesive films are particularly only provided for determination of the reference values, i.e., the known crosslinking degrees, of the reference samples.

[0024] The peroxidic crosslinked polymer may be any polymer, which has undergone a reaction involving the crosslinking of polymers initiated by the formation of peroxide radicals. A notable example of a peroxidic crosslinked polymer is ethylene vinyl acetate (EVA), which is commonly used for the production of photovoltaic modules as an encapsulant for the solar cells.

[0025] A method for the determination of the crosslinking degree of a polymer in a sample is disclosed, the polymer being a peroxidic crosslinked polyethylene copolymer. The method may particularly comprise the following steps:

[0026] - exposing the polymer to near infrared radiation originating from a radiation source, in particular such that the near infrared radiation passes the polymer at least once,

[0027] - determining a set of measurement values being indicative of the absorption of the near infrared radiation by the polymer in a measurement wavelength range, wherein the wavelength range is between 800 nm and 3000 nm,

[0028] - feeding the set of measurement values to a multivariate data model, which multivariate data model is based on a multiplicity of sets of reference values,

[0029] - obtaining, from the multivariate data model, an output value being indicative of the crosslinking degree of the polymer.

[0030] In the multivariate data model, the sets of reference values may be obtained by exposing reference samples to the near infrared radiation; wherein the reference samples comprise or consist of a reference polymer with known crosslinking degree, the reference polymer being a peroxidic crosslinked polyethylene copolymer; and wherein the sets of reference values are indicative of the absorption of the near infrared radiation by the reference polymers in the measurement wavelength range. It may be provided that the sample comprises a cover layer, wherein the cover layer is at least partially transparent for the near infrared radiation, wherein the material of the cover layer is different from the polymer, and wherein the near infrared radiation passes the cover layer when exposing the polymer to the near infrared radiation.

[0031] It may be provided that the cover layer is a glass layer or a polymer layer.

[0032] It may be provided that a reflector is provided, wherein the near infrared radiation is reflected from the reflector after having passed the polymer, such that the near infrared radiation passes the polymer a second time before the set of measurement values is determined.

[0033] It may be provided that the reflector is part of the sample.

[0034] It may be provided that the measurement wavelength range is between 1 ,400 nm and 2,500 nm.

[0035] It may be provided that the sample is a photovoltaic module.

[0036] It may be provided that the multivariate data model is based on a principal component analysis (PCA) or a partial least squares (PLS) regression of the sets of reference values.

[0037] It may be provided that the set of measurement values and the sets of reference values are normalized and / or smoothed.

[0038] It may be provided that the set of measurement values and the sets of reference values are absorption or transmission spectra.

[0039] It may be provided that the set of measurement values and the sets of reference values are n-order derivatives, in particular first-order derivatives, of absorption or transmission spectra. It may be provided that the output value is a classification value, indicating if the crosslinking degree of the polymer is within a pre-determined range of crosslinking degrees.

[0040] It may be provided that the output value is a quantification value, providing a quantitative estimation of the crosslinking degree of the polymer.

[0041] It may be provided that the classification value indicates if the crosslinking degree of the polymer is within a first range of crosslinking degrees or within a second range of crosslinking degrees.

[0042] It may be provided that the polymer is a monolayer or a multilayer.

[0043] Furthermore, a device for the determination of the crosslinking degree of a polymer in a sample is disclosed, the polymer being a peroxidic crosslinked polyethylene copolymer. The device may comprise:

[0044] - a radiation source emitting near infrared radiation towards the polymer, in particular such that the near infrared radiation passes the polymer at least once,

[0045] - a detection unit detecting near infrared radiation, which has been in contact with the polymer, and determining a set of measurement values being indicative of the absorption of the near infrared radiation by the polymer in a measurement wavelength range, wherein the measurement wavelength range is between 800 nm and 3000 nm, and

[0046] - a computing unit, where a multivariate data model is stored, which multivariate data model is based on a multiplicity of sets of reference values,

[0047] - wherein the computing unit calculates, from the multivariate data model, an output value being indicative of the crosslinking degree of the polymer.

[0048] It may be provided that the sets of reference values are obtained by exposing reference samples to the near infrared radiation. It may be provided that the reference samples comprise or consist of a reference polymer with known crosslinking degree, the reference polymer being a peroxidic crosslinked polyethylene copolymer.

[0049] It may be provided that the sets of reference values are indicative of the absorption of the near infrared radiation by the reference polymers in the measurement wavelength range.

[0050] Further features can be derived from the claims, the drawings, and the detailed description of specific embodiments. In the following, exemplary embodiments of the present invention will be described, which serve for illustrative purposes and shall not have a limiting effect on the scope of the claims.

[0051] In the drawings:

[0052] Fig. 1 shows a schematic drawing of an exemplary measurement assembly;

[0053] Fig. 2 shows a graph obtained from the principal component analysis of multiple reference samples;

[0054] Fig. 3 shows another graph obtained from the principal component analysis of multiple reference samples;

[0055] Fig. 4 shows a graph obtained from the partial least squares analysis of multiple reference samples;

[0056] Fig. 5 shows the internal validation result from the partial least squares analysis.

[0057] Except otherwise indicated, the drawings show the following features: Polymer 1 ; sample 2; near infrared radiation 3; radiation source 4; cover layer 5; reflector 6; detection unit 7; photovoltaic layer 8; backing layer 9.

[0058] Fig. 1 shows a schematic drawing of an exemplary measurement assembly, where the sample 1 is a photovoltaic module, which has a multi-layer structure. The outer layer is a cover layer 5 of glass, followed by EVA polymer 1 , the photovoltaic layer 8 in the form of crystalline silicon, another layer of EVA polymer 1 , and a polyester-based backing layer 9. The photovoltaic layer 8 is encapsulated by the polymer 1 . The EVA polymer 1 is a peroxidic crosslinked polymer, the crosslinking degree of which should be determined.

[0059] In Fig. 1 it is shown that from a radiation source 4, near infrared radiation 3 is emitted towards the sample 2. The near infrared radiation penetrates the cover layer 5 and the polymer 1 , before it reaches the photovoltaic layer 8 and is reflected therefrom. The photovoltaic layer 8 therefore acts as a reflector 6 for the measurement. After having hit the reflector 6, the near infrared radiation 3 again passes the polymer 1 and the cover layer 5, before being collected by a detection unit 7. A computing unit will store and process the data read out from the detection unit 7 for further use. A detailed description of an exemplary embodiment of the measurement method is given below.

[0060] Sample preparation

[0061] A range of reference samples, similar to the sample shown in Fig. 1 , was produced in a standardized process, wherein the polymer had different crosslinking degrees. The experimental design followed the industrial practice of controlling the degree of crosslinking mainly via the lamination time while keeping the lamination temperature and the composition of the polymer film constant.

[0062] The un-cured samples were placed in a pre-heated laminator, but were still not in direct contact with the heating plate. Next, the laminator was evacuated for 4 minutes, and the module was shifted to the heating plate kept at a temperature of approx. 150 °C. Upon contact, the pressure in the lamination chamber was reduced to approx. 60 Pa (0.6 mbar) in a second evacuation phase. In a third step, a pressure of approx. 85 kPa (850 mbar) was applied to the stack via a pressure plate, thus initiating the EVA crosslinking phase. For the purpose of this study, the crosslinking time was systematically varied from 10 to 2000 s (10s, 60s, 180s, 300s, 600s, 1200s, 2000s) with the lamination temperature kept constant at approx. 150°C. After the end of this phase, the laminator was vented and the samples were removed. Soxhlet extraction method

[0063] The Soxhlet extraction method is the established “gold standard” used by most module manufacturers to control the lamination quality of EVA encapsulants of PV modules. It was performed in accordance to the optimized procedure from the IEC62788-1-6:2017 standard on “Measurement procedures for materials used in photovoltaic modules - Part 1-6: Encapsulants - Test methods for determining the degree of cure in Ethylene-Vinyl Acetate”.

[0064] In a first step, specimens weighing approx. 0.3 g were cut from different sections of the prepared reference samples. The exact initial weight of each specimen (Mi) was determined on a precision balance. The specimen was then put in a Soxhlet extraction sleeve and extracted for 18 hours in a xylene isomer mixture. After this treatment, the non-crosslinked fraction was supposed to be fully dissolved in the solvent and could be separated from the remaining, crosslinked and hence insoluble, elastomer matrix (“gel”). This insoluble residue stayed in the extraction sleeve and was dried at 60°C in vacuum for 24 hours, followed by the determination of its net weight (M2). The ratio of the mass of the insoluble residue divided by the initial mass of the specimen yields the “gel content”, corresponding to the crosslinking degree:

[0065] Near infrared spectroscopy

[0066] Near infrared (NIR) spectra were measured with a Trinamix handheld NIR spectrometer. For each reference sample (7.e. , crosslinking time) five spectra were measured in transflectance mode between 1400 and 2500 nm. Data processing and multivariate data analysis (MVDA) were performed with the Unscrambler X 10.5 (CAMO Software AS., Oslo, Norway) program. Data modelling - Classification

[0067] Prior to multivariate data analysis and chemometrics the collected sets of reference values, i.e., the NIR spectra in the 1454-2198 nm range, were subjected to normalization applying standard normal variate (SNV) transformation. Each spectrum treated as described here, was taken as a set of reference values within the meaning of the general description above.

[0068] Subsequently principal component analysis (PCA) applying singular value decomposition algorithm, full cross validation with 35 segments (i.e., number of spectra) and 7 components were performed. It was found that two principal components were sufficient to describe more than 99 % of the explained variance in the PCA model.

[0069] Data modelling - Quantitative analysis and prediction

[0070] The same near infrared spectra as described above were also used for quantitative analysis. As spectral pretreatments, SNV transformation was performed and the first derivative of each spectra was calculated in the wavelength range between 1598 and 1902 nm. For the first order Savitzky-Golay derivation a second order polynomial with total of seven (three left and right) smoothing points was used. Each spectrum treated as described here, was taken as a set of reference values within the meaning of the general description above.

[0071] The quantitative analysis was performed by means of partial least squares (PLS) regression applying Kernel algorithm, full cross validation with 35 segments and 7 factors. In this PLS model, 3 factors were necessary to describe at least 99 % of explained variance.

[0072] Results

[0073] Table 1 summarizes the results from the Soxhlet extraction of the reference samples with different crosslinking times. It can be seen that generally with increasing crosslinking time the Soxhlet degree of crosslinking increased as well. Table 1: Soxhlet values of reference samples i ir Crosslinking degree

[0074] Crosslinking time Sample ID (Soxhlet)

[0075] 10 s EVA-GB-v10s 2,30%

[0076] 60 s EVA-GB-v60s 2,60%

[0077] 180 s EVA-GB-v180s 2,00%

[0078] 300 s EVA-GB-v300s 0,50%

[0079] 600 s EVA-GB-v600s 11 ,80%

[0080] 1200 s EVA-GB-v1200s 63,40%

[0081] 2000 s EVA-GB-v2000s 84,70%

[0082] Scatter plots of the PCA model derived from the measurements of the reference samples and their crosslinking degree according to Table 1 are shown in Figs. 2 and 3. They indicate that two clearly separated groups can be identified based on the NIR data. In one group only samples with the two highest cross-linking degrees (Sample IDs EVA-GB-v1200s and EVA-GB-v2000s) can be found, whereas in the other group samples with lower than 15% crosslinking degree are located. This points out that NIR data can be used to classify EVA samples according to their crosslinking degree.

[0083] These classifications can be seen in the scatter plots of principal components 1 and 2 (Fig. 2) as well as principal components 2 and 3 (Fig. 3), which show an even clearer grouping of the two reference samples with the highest crosslinking degrees,

[0084] In order to perform quantitative analysis and prediction a smaller spectral range and the first derivatives of the NIR data was selected. As a result, a highly stable and robust PLS model could be calculated, which applies only three factors to describe the correlation between NIR data and cross-linking degrees. The low number of factors indicates that the model is neither under- or over-fitted. Furthermore, the additional third component compared to the PCA classification provides further chemical information and enables an accurate prediction of cross-linking degree. In fact, the correlation coefficient between the reference and predicted crosslinking degrees for the PLS calibration and validation is above 0.98.

[0085] Fig. 4 shows a scatter plot of score values of the first and second factors. Fig. 5 shows and scatter plot of reference measurements (Soxhlet) versus predicted crosslinking degrees, i.e., an internal validation of the model. The results clearly show that NIR spectroscopy can be used for the determination of the degree of crosslinking of solar cell encapsulants and other peroxidic crosslinked polymers. A qualitative classification approach allows distinguishing between low and high degree of crosslinking samples. Using gel content measurements from Soxhlet extraction, a prediction of the degree of crosslinking from the NIR data was possible. In both cases stable and robust data models with a good accuracy were obtained.

Claims

Claims1. A method for the determination of the crosslinking degree of a polymer (1 ) in a sample (2), the polymer (1 ) being a peroxidic crosslinked polyethylene copolymer, and the method comprising the following steps: a. exposing the polymer (1 ) to near infrared radiation (3) originating from a radiation source (4), in particular such that the near infrared radiation (3) passes the polymer (1 ) at least once, b. determining a set of measurement values being indicative of the absorption of the near infrared radiation (3) by the polymer (1 ) in a measurement wavelength range, wherein the wavelength range is between 800 nm and 3000 nm, c. feeding the set of measurement values to a multivariate data model, which multivariate data model is based on a multiplicity of sets of reference values,- wherein the sets of reference values are obtained by exposing reference samples to the near infrared radiation (3),- wherein the reference samples comprise or consist of a reference polymer with known crosslinking degree, the reference polymer being a peroxidic crosslinked polyethylene copolymer, and- wherein the sets of reference values are indicative of the absorption of the near infrared radiation (3) by the reference polymers in the measurement wavelength range, d. obtaining, from the multivariate data model, an output value being indicative of the crosslinking degree of the polymer (1 ).

2. The method according to claim 1 , wherein the sample (2) comprises a cover layer (5), wherein the cover layer (5) is at least partially transparent for the near infrared radiation (3), wherein the material of the cover layer (5) is different from the polymer (1 ), and wherein the near infrared radiation (3) passes the cover layer (5) when exposing the polymer (1 ) to the near infrared radiation (3).

3. The method according to claim 2, wherein the cover layer (5) is a glass layer or a polymer layer.

4. The method according to any of claims 1 to 3, wherein a reflector (6) is provided, wherein the near infrared radiation (3) is reflected from the reflector (6) after having passed the polymer (1 ), such that the near infrared radiation (3) passes the polymer (1 ) a second time before the set of measurement values is determined.

5. The method according to claim 4, wherein the reflector (6) is part of the sample (2).

6. The method according to any of claims 1 to 5, wherein the measurement wavelength range is between 1 ,400 nm and 2,500 nm.

7. The method according to any of claims 1 to 6, wherein the sample (2) is a photovoltaic module.

8. The method according to any of claims 1 to 7, wherein the multivariate data model is based on a principal component analysis (PCA) or a partial least squares (PLS) regression of the sets of reference values.

9. The method according to any of claims 1 to 8, wherein the set of measurement values and the sets of reference values are normalized and / or smoothed.

10. The method according to any of claims 1 to 9,- wherein the set of measurement values and the sets of reference values are absorption or transmission spectra, or- wherein the set of measurement values and the sets of reference values are n- order derivatives, in particular first-order derivatives, of absorption or transmission spectra.11 . The method according to any of claims 1 to 10,- wherein the output value is a classification value, indicating if the crosslinking degree of the polymer (1 ) is within a pre-determined range of crosslinking degrees, or- wherein the output value is a quantification value, providing a quantitative estimation of the crosslinking degree of the polymer (1 ).

12. The method according to claim 11 , wherein the classification value indicates if the crosslinking degree of the polymer (1 ) is within a first range of crosslinking degrees or within a second range of crosslinking degrees.

13. The method according to any of claims 1 to 12, wherein the polymer (1 ) is a monolayer or a multilayer.

14. A device for the determination of the crosslinking degree of a polymer (1 ) in a sample (2), the polymer (1 ) being a peroxidic crosslinked polyethylene copolymer, the device comprising:- a radiation source (4) emitting near infrared radiation (3) towards the polymer (1 ), in particular such that the near infrared radiation (3) passes the polymer (1 ) at least once,- a detection unit (7) detecting near infrared radiation (3), which has been in contact with the polymer (1 ), and determining a set of measurement values being indicative of the absorption of the near infrared radiation (3) by the polymer (1 ) in a measurement wavelength range, wherein the measurement wavelength range is between 800 nm and 3000 nm, and- a computing unit, where a multivariate data model is stored, which multivariate data model is based on a multiplicity of sets of reference values,- wherein the sets of reference values are obtained by exposing reference samples to the near infrared radiation (3),- wherein the reference samples comprise or consist of a reference polymer with known crosslinking degree, the reference polymer being a peroxidic crosslinked polyethylene copolymer, and- wherein the sets of reference values are indicative of the absorption of the near infrared radiation (3) by the reference polymers in the measurement wavelength range, wherein the computing unit calculates, from the multivariate data model, an output value being indicative of the crosslinking degree of the polymer (1 ).

Images