O. Daikos, M. Naumann, K. Ohndorf, C. Bundesmann, U. Helmstedt, T. Scherzer
Talanta 223 (2021) 121696
10.1016/j.talanta.2020.121696

The thickness of thin layers of the conductive polymer PEDOT:PSS in the range bet­ween about 60 and 300 nm was determined by a near-infrared spectroscopic method using a hyperspectral camera. The reflection spectra of the layers do not con­tain bands, but consist of a moderate slope of the overall reflect­ance in the range between 1320 and 1850 nm. Despite the low thick­ness, the spectra show an extremely strong dependence on the thickness of the layers, which allows their use for quan­titative measurements. The prediction of quantitative thick­ness data from the reflection spectra was based on a chemometric approach using the partial least squares (PLS) algorithm. Calibration was carried out by means of spin-coated layers of PEDOT:PSS, whose thickness was determined by white-light inter­ferometry and sty­lus profi­lo­metry. Finally, this resulted in a calibration model with a root mean square error of pre­diction (RMSEP) of about 9 nm. After external validation of this model, it was used for quantitative imaging of the thickness distribution in PEDOT:PSS layers. The precision of the predicted values was confirmed by comparison with data from the reference methods. Moreover, it was shown that this approach can be also used for hyperspectral imaging of the thickness of thin printed layers and structures of this conductive polymer on polymer film or paper with excellent thickness resolution. This analytical approach opens new possibilities for in-line process con­trol by large-scale monitoring of thickness and homogeneity of thin layers of conductive polymers.

O. Daikos, T. Scherzer
Talanta 221 (2021) 121567
10.1016/j.talanta.2020.121567

Hyperspectral imaging was used for large-scale monitoring of the residual moisture in wide tex­tile webs at the end of the drying process that follows their washing or finishing by impreg­na­tion in aqueous solutions or dispersions. Such data are essential for optimizing the energy effi­ciency and the precise control of the dry­ing process. Quantitative ana­ly­sis of the re­corded spec­tral data was carried out with multi­variate regression methods such as the partial least squares (PLS) algorithm. Reference data for calibration of the prediction models were de­ter­mined by gravimetry. The drying of textile ma­terials from both natural or synthetic fibers pos­ses­sing different water absorption capacities (cot­ton, polyamide, polyester), which were par­tial­ly finished with an optical brightener, was in­ve­stigated. Moisture contents in the range from 0 to about 12 wt% were considered in the ca­li­bration models. For all systems, the root mean square error of pre­diction (RMSEP) for the re­sidual moisture was found to be about 0.5 wt%, that is, about 1 g/m². In addition to the quan­ti­tative determination of the water content, hyper­spectral imaging provides detailed information about its spatial distribution across the textile web, which may help to improve the control of the drying process. In particular, it was demon­strated that the developed methods were ca­pable of detecting and visualizing inhomogeneous moist­ure distributions. Averaging of the in­di­vid­ual values of the moisture content pre­dicted from all spectra across the surface of the tex­tile samples resulted in a very close correlation with the cor­re­sponding gravimetric re­fer­ence values. Due to the averaging process, the difference bet­ween both values is generally lower than RMSEP even in case of samples with inhomo­geneous distribution of the moisture. The high precision and the broad capabilities of the developed analytic methods for in-line mo­nitoring of the moisture content hold the potential for an efficient process control in tech­nical textile converting processes.

O. Daikos, K. Heymann, T. Scherzer
Prog. Org. Coat. 132 (2019) 116
10.1016/j.porgcoat.2019.03.008

Near-infrared (NIR) chemical imaging was used for in-line monitoring of the conversion in thick UV-cured white-pigmented acrylate coatings applied to various substrates such as glass, stain­less steel, PVC and glass fiber reinforced plastic (GRP) boards. Quantitative results were ob­tained by means of chemometric calibration models based on the partial least square (PLS) al­go­rithm. Two spectroscopic techniques were tested for their potential for the charac­te­ri­zation of the con­version after UV irradiation with a broad range of UV doses in order to provide reliably precise re­ference data for the calibration. NIR reflection spectroscopy using band integration of the acryl­ate band at 1620 nm was selected as method of choice. Using glass as substrate for the coatings at the beginning, a PLS model was established and evaluated for its performance to predict the con­version in independent samples. The acrylate conversion was predicted with an error of 2 %. Ge­nerally, any calibration model is specific to a well-defined sample system (e.g. material and thickness of substrate and coat­ing). In order to reduce the effort required for the development of spe­ci­fic calibration models for each substrate used in this study, methods for calibration transfer were de­veloped. It was found that the requirements for such transfer depend on the optical pro­per­ties of the substrate. In case of stainless steel and PVC boards, simple preprocessing of the spec­tra by base­line correction and normalization (similar to samples on glass) led to sa­tis­fac­to­ry results, which is confirmed by prediction of the conversion with similar error margins as for coat­ings on glass. In case of glass fiber boards, the spectrum of the pristine GRP board had to be sub­tracted from the spectrum of the coated sample before the transfer of the PLS model. The re­sult­ing prediction er­ror (RMSEP) was found to be 3.6 %. The comparison between this trans­ferred and a spe­cific PLS_GRP mo­del that was built up for evaluation only proved the high per­form­ance of the trans­ferred PLS mo­del. These results clearly demonstrate the high efficiency of the trans­fer of spe­cific PLS mo­dels to dif­fer­ent (but similar) sample systems such as coatings applied to other sub­strates. In addition to the quan­titative determination of the conversion, the developed cali­bra­tion models were also used for the evaluation of its spa­tial di­stri­bution across the surface of UV-cured coatings.

G. Mirschel, O. Daikos, T. Scherzer, C. Steckert
Talanta 188 (2018) 91
10.1016/j.talanta.2018.05.050

This paper demonstrates for the first time that near-infrared (NIR) chemical imaging can be used for in-line analysis of textile finishing processes based on impregnation. In particular, it was shown that this analytical me­thod is sufficiently sensitive for the quantitative deter­mi­na­tion of the application weight of rather thin layers of fi­nishing chemicals. Quantitative ana­ly­sis of the data recorded by a hyper­spec­tral ca­mera (1320-1900 nm) was based on chemo­metric approaches using the partial least squares (PLS) algorithm. In this work, a flame retardant and a polyvinyl acetate-based stiffe­ning agent applied to poly­ester or cotton fabrics, re­spec­tive­ly, were stu­died with application weights in the range between about 1 and 50 g m^-2. For both sy­stems, the prediction error (RMSEP) was found to be about 1.5 to 2 g m^-2. Averaging of the pre­dicted individual values of the application weight of the finishes across the complete surface of the fabric resulted in a very close correlation with the cor­re­sponding reference values ob­tained by gra­vi­metry. Furthermore, NIR chemical imaging was used for the detec­tion of re­maining traces of a size (a processing agent) after washing, which had to be washed-out be­fore sub­sequent pro­ces­sing steps. Results of the pre­sent in­ve­stigations prove that even for very thin size layers bet­ween 0.4 and 5.5 g m^-2 the ap­pli­ca­tion weight can be predicted with a precision of about 0.4 g m^-2.

Apart from the quantitative determination of the application weights, the use of NIR chemical imaging for the analysis of finished textiles was mainly directed towards the investigation of the spatial distribution or the homogeneity of the applied colorless finishes across the surface of the fabrics. It was shown that this method is able to detect and visualize various inhomo­gene­ities on the finished textiles resulting for instance from processing defects or from vari­ous tech­nical effects that may influence the drying pro­cess and con­se­quent­ly the spa­tial di­stri­­bu­tion of the finish. Moreover, the distribution of traces of size that had been sprayed pur­pose­ly on a washed polyester fabric could be detected.

All measurements in the present study were carried un­der conditions that were very similar to those in typical technical processes (e.g. with respect to line speed). Therefore, the out­stand­ing performance of the method opens an immense potential for application in process and quality control.

Gabriele Mirschel, Olesya Daikos, Tom Scherzer, Carsten Steckert
Anal. Chim. Acta 932 (2016) 69-79
https://doi.org/10.1016/j.aca.2016.05.015

This paper demonstrates for the first time that near-infrared (NIR) chemical imaging can be used for in-line analysis of textile lamination processes. In particular, it was applied for the quan­ti­tative determination of the applied coating weight and for monitoring of the spa­tial di­stri­bution of hot melt adhesive layers using chemometric approaches for spectra evaluation. Layers with coat­ing weights between about 25 and 130 g m^-2 were used for the lamination of poly­ester fabrics and nonwovens as well as for poly­urethane foam. It was shown that quan­ti­ta­tive data with adequate precision can be actually ob­­tained for layers applied to materials with sig­nificantly he­te­ro­geneous surface struc­ture such as foam or for hidden layers inside fabric la­minates. Even the coating weight and the ho­mo­geneity of adhesive layers in com­po­sites con­sist­ing of black textiles only could be quan­ti­ta­tively analyzed. The pre­diction er­rors (RMSEP) determined in an external validation of each calibration model were found to range from about 2 g m^-2 to 6 g m^-2 depending on the specific system under investigation. All ca­li­bra­tion models were applied for chemical imaging in order to prove their per­for­mance for mo­ni­tor­ing the thickness and the homogeneity of adhesive layers in the vari­ous textile sy­stems. More­over, they were used for the de­tection of irregularities and coat­ing defects. Inve­sti­ga­tions were carried out with a large hyper­spec­tral ca­mera mounted above a conveyor. There­fore, this method allows large-area monitoring of the properties of laminar ma­ter­ials. Con­se­quently, it is potentially suited for process and quality control during the lamination of fab­rics, foams and other materials in field-scale.